Abstract: In view of problem of bad effect of image collected by industrial endoscope because of special application place, the paper proposed an image fusion algorithm based on multi-scale wavelet transform and different fusion operators. Firstly, it made wavelet transform for source image, then used simple average algorithm to fuse low frequency coefficients which represents approximate information and used dynamic weighting average algorithm based on local deviation to fuse high frequency coefficients which represent detail information. Finally, it made inverse wavelet transform of processed low and high frequency coefficients to get fused image. The experiment result showed that the algorithm improves image processing effect with better image parameters comparing with other algorithms.