Support vector machine is the uniform method of statistical learning method,and has become more and more popular in research field.Support vector machine has achieved excellence in pattern recognition and text classification for its high performance in veracity.Support vector machine method was used to process several Binary image and Gray scale image and got a good statistical result.A strategy to select feature vectors has been found out,and the criterion to judge the results are discussed.Compared with other methods,the important conclusion was draw: it performs perfectly in statistic when using support vector machine to detect the image edge.