A fuzzy reliability growth model was presented on the basis of a Bayesian reliability-growth model based on Dirichlet prior distribution using fuzzy theory. The method for determining prior parameters was studied using Dirichlet prior distribution. A metal/water reaction ramjet's reliability growth test data was analyzed using this distribution. The prior distribution and the posterior distribution at each test stage were presented. And the effect of the parameter related to prior information's variance on posterior reliability was analyzed. Fuzzy variable was introduced in the original model to build a fuzzy reliability growth model. Using the model, the fuzzy reliability at each stage was calculated.