As the source-load coordination in the dispatch optimization of island micro-grid is very important but often neglected, based on the operation characteristics of island micro-grid and load classified demand response, a multi-objective optimal dispatch model considering source-load coordination of micro-grid with goals of minimizing operation cost and pollution emission is established. By introducing multi-core parallel computing environment and multi-population parallel crossover mutation mechanism, a new Parallel Multi-objective Differential Evolution (PMODE) algorithm is designed to solve the model effectively, in order to harmonize the contradiction between optimization depth and speed of conventional intelligent algorithm. Based on the application results, the influence of classified load transfer and pre-and after cutting on the output of each distributed power in microgrid and the load adjustment of peak-valley are analyzed. The results show that the proposed optimal dispatch model can effectively realize energy saving and emission reduction and improve the utilization rate of wind/photovoltaic power, verifying the superior performance of PMODE. This work is supported by National Natural Science Foundation of China (No. 51567007 and No. 51867008) and Natural Science Foundation of Jiangxi Province (No. 20171BAB206042) and Jiangxi Innovate Foundation of Graduate (No. YC2017-S251).