During product platform life cycle, innovation problem identification and decision-making are regarded as vital issues in product platform evolution process. The Comprehensive Disturbance Degree is proposed and analyzed considering customer demand, technology status and production capacities, then existing problems of product platform are identified. According to Innovation Problem Selecting Principles, the innovation problem set Q is defined. The value of modules in product platform is calculated using Value Engineering, the module set M needed to be improved is determined. Then based upon the correlation degree analysis of the innovation problem set and the module set, Fuzzy Clustering Algorithms is advanced to classify innovation problems. Finally, a case study is given to illustrate the validity of the methodology.