Vacuum drying process for rapeseed was optimized by combination of neural network and particle swarm algorithms. BP neural network algorithm was used to establish the three layer network model to forecast the relationship between the drying temperature, initial moisture content, pressure and the average rate of moisture dropping, germination rate of rapeseed. Network weight and threshold of the model was calculate by using the sample data from the experiment. Then, PSO algorithm was used to optimize the initial parameters of the model. It is verified by experiment that the maximum relative error of BP networkmodel was 4.5%, whereas it was 2.93% for PSO–BP network model. The combination of BP neural network and PSO algorithms could decrease the error between the actual value and network simulated value.