Although the role of farmer cooperatives as a social unit can have impact on their performance, empirical analysis on how societal output and social value relevant variables affect the cooperatives performance is sparse. The objective of this paper is to provide an economic framework and operational model for performance measurement of farmer cooperative associated with societal impact. A multi-output translog production function considering social output represented by the number of beneficiary farmers using data from surveys 164 cooperatives in Fujian province, China, is estimated. The average technical efficiency of cooperatives is estimated to be 0.747, implying that cooperatives can be increased by 25.30% without any additional resources given the current production input level. It is interesting to find that cooperatives efficiency scores and their rankings are significantly different with and without taking societal output into account, which indicates that social output created by the number of beneficial farmers cannot be ignored when evaluating cooperative s performance. The societal value relevant variables for technical inefficiency factors represented by extent of providing members service, namely training members and selling products are also found negatively affecting technical efficiency of cooperatives. The findings indicate the evaluation of cooperatives performance should consider their non-economic social contribution. Acknowledgement : The authors gratefully acknowledge financial support from Fujian Agriculture and Forestry University FAFU university: 2016 project (social science category) for supporting outstanding young scientific researchers: (NO: xjq201632).We also acknowledge the kind help from Professor Jerker Nilsson.