A novel adaptive image enhanced fusion algorithm is presented. The new algorithm separates original images into low frequency and high frequency components. The low frequency component represents the local luminance mean feature, and the high frequency component represents local contrast feature. Using enhance operator based on local gray variance, and nonlinear mapping operator, the algorithm performs adaptive enhancement of infrared and visible images. The feature is extracted and fused image is constructed based on local spatial energy and normalized correlation measure within neighborhood and weighted average fusion operator. The salient detail feature information of the input images is presented in the fused image.