To overcome the shortcomings of entropy bas ed multilevel segmentation methods, such as high computational complexity and po or segmentation performance, a fast multilevel thresholding for image segmentation method based o n generalized probability Tsallis entropy is proposed. First the t raditional gray probability is modified into generalized probability to form gene ralized probability Tsallis entropy (GPTE). Then the parameters of GPTE are cho sen adaptively through the image histogram average value. Thirdly， a multilevel t hresholding method based on GPTE is formulated to get more effective segmentation. Finally the differential evolution(DE) and the recursive algorithm are combined and introduced into the multilevel thresholding method to find the best threshold vector quickly. Experi mental results of image segmentation show that the propos method can obtain bett er segmentation results and adaptability with less computation time compared with the tra ditional entropy based multilevel segmentation methods.