Abstract—Establishing innovation alliance has now become a very important and indispensable way for achieving the mission of research and development (R&D) of national mega projects in China. As an important component for constructing the collaborative partnership system, an appropriate approach of partner selection should be able to evaluate the performance of candidates objectively. Innovation alliance has members with different backgrounds, but the most suitable partner should be involved in the alliance by launcher before establishing alliance. For this reason, the aim of this paper is to take the different characteristics of organization as the influence factors into the account of weight setting, and propose a fuzzy analytic hierarchy approach (fuzzy AHP) to effectively evaluate candidates. According to the extension principle of fuzzy set theory, linguistic variables defined as fuzzy numbers are applied to pair-wise comparisons to avoid the vague situation, and this study proposes an approximate method for calculating the multiplication products of fuzzy number. This method can handle the vagueness and incompletion during the process of evaluation. A case study is also given to demonstrate the potential of the methodology.