To obtain excellent comprehensive performance for mechanical product, it’s necessary to select the best one from multiple design schemes. Scheme criteria evaluation matrix was standardized and criteria weight vector of every design scheme was optimized. According to the principle of minimum offset of design scheme evaluation, scheme evaluation problem was linearly programmed. Uncertainty factors of criteria weight were analyzed, and entropy was introduced to describe the uncertainties. According to the maximum entropy theory, design scheme evaluation model was built, which was solved by building Lagrange function and forming nonlinear equations. Finally, these design schemes could be sorted to obtain the best one. The example demonstrated validity of the proposed evaluation method.