| Abstract | Most of the resource-poor farmers in Thailand are facing water scarcity for cultivation, especially in the rainfed condition. This situation reduces agricultural yields and the farmer household economic. In Thailand, attempting to mitigate poverty has been a failure for satisfactory of the goal because of focusing in an activation of macro-economic, whereas the micro-economic as the household farming level has been left behind. However, experiences of the literatures have proven the agricultural planning, which cope with technology efficiency, economics, finance, social and environment, as the key success in the mitigation of poverty, and rise up the rural household economic. As the new modern technologies in agriculture rapidly increase and accrue several benefits to farmers, there are the changes in Thai agricultural farm operations, its input resources management, and environment quality. These strongly influence an adverse environment by the deterioration of the water and land resources, and the contribution of a substantial global warming through an increase of greenhouse gas (GHG) emission. The appearance of complexity in agricultural planning decision has been generated by such situations as rural Thai agriculture paradigms involve with the integrated agriculture productions systems (IAPSs). This research deals with developing MCDM for formulating IAPSs and taking into account the energy and environmental constraints for the study area. As assessment of the present pattern of energy consumption of farm operations and energy resources utilization by different agricultural productions and different IAPSs in the study area were analyzed, the corresponding GHG as equivalent carbon dioxide (CO 2 e) was also evaluated in terms of energy resource inputs utilization and fossil-fuel consumption by farm operations. To construct the MCDM framework, two core decision processes as a glance-based decision process (GBDP) and an optimization-based decision process (OBDP) were developed as a simple and a higher complex decision. The MCDM was focused at the micro-level (i.e. field plot). OBDP required multi-objective evolutionary algorithm (MOEA) in the simulation-based optimization process. Two MOEAs, namely elitist multi-objective evolutionary algorithm (NSGA II), and multi-objective differential evolution (MODE) were comparatively implemented for integrating into the MCDM framework. 281 farm plots were surveyed to collect important information for analyses. Through the energy analyses, 4 indicators were applied for an assessment of agricultural productions’ energy efficiency. Information of yield productions, income, and net revenue of the productions within the study area also were intensively analyzed. The MODE algorithm was evaluated against the original NSGA II using the eleven standard multi-objective test functions. GBDP applied the AHP approach to obtain the subjective decision of DMs related to IAPSs decision and agricultural planning. 83 DMs were asked of willingness for assigning weighted importance values of dimensions and sub-criteria. 12 sub-criteria were constructed under 4 main dimensions (physical, economics, environment, and social culture). OBDP considered the optimal allocation of crops area for modified IAPSs in three LHCs according to the eleven objectives designed. The modified IAPSs considered the introduction of soybean, vegetables, and pond-culture to the systems. OBDP incorporated the simulation model, probabilistic structure of rainfall occurrence, and the stochastic approach in the optimization. A Stochastic Multi-State First Order Markov Chain (SMFOMC) was developed for synthesizing daily rainfall sequence. Six probability distribution functions were studied to classify three water years, namely the less water year, the normal water year, and the much water year. Example integration of livestock production was also performed for evaluation. The findings indicated the significant variations among IAPSs in their energy input-output characteristics. For crops, the maximum energy consumer was cassava (32.4 GJ/ha) with its great seed energy resource. The total energy input of other crops varied in the range of 20-54% of that of cassava. Operation energy of crops was in the range of 3.5-9.5 GJ/ha, which mostly derived from tillage/ploughing and the transportation operations; planting, irrigation, and transportation for vegetables were predominant, which constituted energy input to higher than 10 GJ/ha. Vegetables, mostly, gained EER lower than one except chilies. Soybean, fruit (banana), and ruzi grass contributed EER higher than 10, whereas, maize received the maximum EER of 40, which mostly relied on its energy of by-products. Paddy rice and cassava contributed EER in the range of 4.1-6.5. Fossil fuel received the highest level dependency in crops productions, including pond-culture, which translates into the largest emission of CO 2 e as compared with other energy inputs. Cassava employed large amount of fossil energy in harvesting, whereas vegetables, especially chilies, spent it for irrigation and transportation. Transplanted rice provided the highest CO 2 e emission (906 kg CO 2 e/ha) among crops, IAPSs as rice-grass-cassava-pond corresponded to the highest CO 2 e emission. Amidst of global debate that paddy rice cropping is the major source of agricultural GHGs, the current findings confirm the notion; in which almost 50% of its total CO 2 e emission comes from methane (CH 4 ). Ruzi grass, fruit (banana), maize, and soybean did not produce considerable CO 2 e emission compared to paddy rice, cassava and vegetables. CO 2 e emission from pond-culture was also trivial. The findings also showed that GBDP can improve NV of existing IAPSs in the range of 3% to more than 100%, whilst it can increase NV from a business as usual (BAU) case to be more than 1.2 times under scenarios assigned. Furthermore, NV was projected to be 5.26 Million US$ for the whole study area. For OBDP, the findings found that the modified IAPSs can preserve the land area of the paddy for securing the food buffer, which is the most suitable choice for Thai resource-poor farmers. Total NV can shift up the minimum positive net monetary earns from about 700 US$/ha to be more than 1,000 US$/ha. All modified IAPSs contributed to the maximum GHG less than 700 kg CO 2 e/ha in contrast with 1,100 kg CO 2 e/ha for the existing IAPSs. Other objectives also contributed to the significant improvement. Pond-culture met a success for practical serving the water needs for the candidate time of the supplementary irrigation. Although the result of pilot example integration of livestock production showed failure for an improvement of NV for the modified IAPS; however, that case was trivial because of a small drop in NV only 0.6%, as it can sustain the environment and energy conservation if GHG and energy input of the livestock were not accounted. In addition, the optimal modified IAPSs effectively contributed to the water productivity ratio in the range of 1.13-2.83. Through the MCDM framework, GBDP and OBDP provided significant improvement for Thai integrated agricultural production systems in Thai rainfed condition according to the agricultural sustainability concepts. |