During the operation of nuclear reactors, zirconium alloy claddings are subjected to significant cyclic stresses, necessitating a comprehensive assessment of the material’s fatigue properties to prevent structural failure. This study focuses on enhancing the accuracy of predictive models for the fatigue performance of zirconium alloys by employing fatigue test data from alloys in two distinct heat-treated states. A three-parameter Weibull distribution model, formulated using the probabilistic weighted moments method, was developed to construct Reliability-Stress-Number of cycles (R-S-N) curves. These curves were rigorously compared and validated against the R-S-N curves obtained using the traditional Basquin model. The findings indicate that the S-N curves, derived from the three-parameter Weibull distribution, demonstrate superior fitting accuracy and markedly surpass the performance of those obtained through conventional methodologies. This advanced modeling approach thus proves to be highly effective and applicable for the detailed analysis and prediction of fatigue properties in zirconium alloys, offering significant implications for the design and safety of nuclear reactors.