During the operation of nuclear reactors, zirconium alloy claddings are subjected to significant cyclic stresses, necessitating an accurate evaluation of their fatigue properties essential for preventing structural failure. This study proposes a high-accuracy prediction method for the fatigue life of zirconium alloys by employing fatigue test data from alloys subjected to two distinct heat-treatment conditions. A three-parameter Weibull distribution model was established using the probabilistic weighted moments method to construct reliability-stress-number (R-S-N) cycles curves. These curves were rigorously compared and validated against those obtained from the traditional Basquin model. The findings indicate that the R-S-N curves derived from the three-parameter Weibull distribution demonstrate superior fitting accuracy and significantly outperform the conventional model. This advanced modeling approach provides a reliable and effective means for predicting the fatigue behavior of zirconium alloys, offering significant implications for the structural design and safety assessment of nuclear reactors.