In this paper, the NN (Neural Network) and SVR (Support Vector Regression) algorithms were discussed and proposed, based on NN and SVR algorithms, for the purpose of solving the over-supply of off-season longan. Integrating Fuzzy prediction theory with NN and SVR provided more accurate and precise results and significantly saved on the computational runtime required during training. Unfortunately, the reducing input models either RINN or RISVR based was illustrated to save computation runtime but obtained low accurate and precise results. For further study, their optimal structure may be the problem interested which can be introduced particle swam search approach.