The environmental effect evaluation of agricultural projects is current a new research field. With the increase of projects, the study of environmental effect appears more and more important. We applied PSO-BP to evaluate the environmental effect of an agricultural project. Emulation and experiment show that the method of neural network algorithm trained by particle swarm optimization has overcome its disadvantage of slow convergent speed and shortcoming of local optimum. Particle swarm optimization needs only a few parameters and is a simple, while the result is pretty good.