Considering the traditional image fusion methods easily reduce the contrast and sharpness of image, blur edge and loss details, a fusion method of infrared and visible images is proposed based on intuitionistic fuzzy set and regional contrast in the fusion framework of non-subsampled contourlet transform. Firstly, the high and low frequency components of source images are obtained by using non-subsampled contourlet transform. Then the low frequency components are combined by intuitionistic fuzzy set including double-Gaussian function due to the characteristic that intuitionistic fuzzy set can describe the fuzzy concept flexibly and accurately. The high frequency components are combined by regional features contrast and Euclidean distance method because of the feature that the method can describe the image texture information in detail. Finally the fusion image is obtained by performing inverse non-subsampled contourlet transform. The experimental results show that the proposed fusion method outperforms traditional methods which deepens the contrast of images, retains the edge and detail information in the source image, and has a better evaluation value.