Aiming at the time consuming and computing complexity of the traditional mean shift segment methods，an adaptive mean shift segmentation algorithm is presented for image segmentation. The algorithm uses a histogram of the image to estimate the probability density of the image. According to the characteristics and probability distribution of each pixel its bandwidth is calculated. Experimental results show that the improved algorithm is a good solution to the problem of low efficiency bandwidth determination and poor segmentation of the fixed bandwidth mean shift algorithm. Adaptive mean shift segmentation algorithm can have very good segmentation results.