Rotary Stirling coolers are critical components in missile-borne infrared thermal imaging systems. Accurately predicting their remaining life after long-term storage is essential for ensuring missile precision and operational effectiveness. This study aims to develop a reliable method for estimating the remaining life of rotary Stirling coolers that have been in storage for over ten years.Based on an analysis of storage failure mechanisms, the study considered factors such as working environment and operating conditions. Accelerated tests were designed to simulate two primary failure mechanisms: working gas pollution and leakage. Samples with over ten years of application history underwent high-temperature storage tests at 80 °C and temperature cycle tests between -55 °C and 80 °C. Storage life prediction models were established for each failure mechanism. Failure analysis was conducted on samples that failed during testing.High-temperature storage tests revealed that gas pollution was not the primary cause of failure, as samples did not fail after 5 760 hours of accelerated testing (equivalent to 14.2 years under normal conditions). Temperature cycling tests proved more effective in inducing failures. Two samples failed after 330 and 270 cycles, respectively, due to excessive leakage rates (Tab.2). Failure analysis using scanning electron microscopy revealed cracking between the metal substrate and coating layer in one sample (Fig.3), and partial extension and loss of the outer coating layer in another (Fig.4). These findings suggest that temperature cycling has a more significant impact on cooler storage life than gas pollution. A prediction model based on the Norris-Landzberg equation was proposed for estimating remaining life under temperature cycling conditions.This study provides valuable insights into the long-term storage behavior of rotary Stirling coolers. While working gas pollution appears to have a slower impact on storage life, gas leakage induced by temperature cycling emerged as the dominant failure mechanism. The developed prediction models offer a foundation for estimating remaining cooler life. However, due to limited sample size and potential environmental prediction biases, further research is needed. Future work should focus on increasing sample size, refining temperature cycling test parameters, and incorporating more field usage data to enhance the accuracy and practical significance of the remaining life prediction models.