Inherent ambiguity and uncertainty in the nature of human resource because of bounded rationality and cognitive limitations always make difficult to predict behaviors of such complex system. Thus, predicting in this area necessitates modeling approaches to model ambiguity as part of system. The purpose of this study is to apply artificial intelligence and advance optimization algorithms to modeling personnel efficiency. To do so, it uses “Emotional Quotient (EQ)” and “individual characteristics” as input variables and “responsibility”, “work speed” and “work accuracy” as output variables. In order to model personnel efficiency, an Adaptive Neuro-Fuzzy Inference Optimized System (ANFIS) is introduced. This system utilizes genetic algorithm and Singular Value Decomposition (SVD) method. It can predict personnel efficiency with minimum training error, minimum predicting error and maximum adaptability to the real efficiency. It is worth mentioning that, for 84% to 96% of records, the extracted models are the same as the real efficiency.