In order to remove the salt and pepper noise in the image and preserve the image detail information, a generalized regression neural network model is proposed, which is suitable for image denoising. Firstly, the principle of the traditional generalized recurrent neural network is analyzed, and the generalized regression neural network is designed. Then, the unique adjustable parameter (smoothing factor) in the generalized regression neural network is optimized. The normalized mean square error and peak SNR are used to analyze the performance of the proposed algorithm. The simulation results show that the proposed algorithm has stronger denoising ability, higher peak signalnoise ratio and lower normalized mean square error compared with RBF neural network and traditional generalized regression neural network, which verifies the effectiveness and advancement of the proposed algorithm.