Aim at solving the existing problems of 3D model retrieval based on neural network, this paper proposes a new algorithm based on BP-bagging. Through bagging, the algorithm turns the weak classifier into the strong. As to feature extraction, the algorithm projections 3D model into six 2D images by six perspective points. Then transforms the images into frequency domain, gets the high dimension feature by two 1D Fourier Transformation, and compress the high dimension feature to the lower, inputs into the BP-bagging classifier to retrieval the 3D model. Proved by the experiment, the new algorithm is more effective in 3D model retrieval.