A leukocyte classification method was proposed by using image technologies.Firstly, based on image color information, image distance transformation and the Snake of Gradient Vector Flow (GVF Snake), the leukocytes were extracted in a blood cell image, and then the high saturation trait of the leukocyte nuclei was combined the morphological mathematics and GVF Snake to detect the nuclei in the leukocyte image. According to the features of morphometry, color and texture for cells, the Support Vector Machines (SVMs) were taken to classify the leukocytes. The results show that the proposed image segmentation method and the classifier to classify the leukocytes can achieve the accuracy by 89.6%. Compared to other traditional cell image segmentation and analysis methods, the proposed method is satisfactory.