China will develop into an innovative country in 2020. It has become an important topic that study on evaluation method of innovation ability. But there are many human factors in most evaluation methods. This would diminish the validity of a systematic review. The unsupervised clustering algorithm, combining fuzzy C means clustering and support vector machine are used in this paper. The impact of human factors will be overcome. Through the 2013 yearbook data analysis of the experimental results, this method is achieved very good results in evaluation of regional innovation capacity.