A mechanized grain loading process for high & large warehouse was provided in order to solve some common problems such as difficulty in moving loading machine, poor efficiency and less loading volume. The optimization of related parameters of loading process were analyzed through field research in grain depots of Sinograin. The mathematical model of multi-objective optimization for the process was proposed according to loading volume, grain surface roughness and number of working position. Based on ant colony algorithm, optimal algorithm flowchart and its relative parameters were defined. Finally, loading process simulations were performed by using MATLAB tools for programming to get the optimal process of movable grain loading machine aiming at the warehouse with size of 30×24 m2. The simulation results showed that the optimal process conformed to manual loading. Loading volume and grain surface roughness were the same with actual loading results by using the optimal process, which reduced the number of loading working position and movement of loading machine, therefore improved working efficiency and reduce labor intensity.