Risk response action (RRA) selection plays an important role in project risk management (PRM). In this paper, a method based on case-based analysis and fuzzy optimization model is proposed in order to help project manager find out available historical RRAs, and determine the optimal set of the RRAs. The steps of the method include: a) representing the target cases and the historical cases; b) determining the fuzzy similarities between the historical cases and target cases; c) screening out the historical cases with the fuzzy similarities higher than the threshold; d) revising the alternative historical RRAs in accordance with the current project; e) determining the optimal set of RRAs with optimization model. The computation results of the method based on the analysis of a case highlight two managerial implications. First, to perform better risk response in future, organizations should always capture a long-term perspective, with an awareness of keeping documents of all handled historical cases. Second, since each RRA obtained from historical cases needs to be adapted in accordance with current situation, adaptation costs should also be taken into account when allocating budget for implementing the selected RRA set.