The image segmentation is a major aspect in image processing, it will produce influences on the image analysis and object recognition. The thresholding segmentation is one of image segmentation methods whose task is to search for the optimal threshold used by segmentation. In this paper, an image segmentation approach using evolution strategies technique is presented. The evolution strategies use mainly the selection and mutation operations to realize the evolutions of populations. In order to improve the efficiency of the approach, the crossover operations among a part of individuals in population and the age parameter of an individual are introduced into the basic evolution strategies. Some experimental results show the proposed approach can get the near optimal threshold being used to image segmentation.