A time-varying reliability analysis method (iTRPD) based on multi-peak distribution is proposed to address the engineering problem of mechanism reliability, with this method applied to the reliability analysis of largescale morphing wing mechanism. First, the morphing wing structural model is discretized into several instantaneous functional functions and converted into independent normal variables. Then, the instantaneous reliabilities at different moments are calculated, while the autocorrelation coefficient matrix between each vector is calculated to get the corresponding probability density function. Finally, according to the covariance property and the correlation between the vectors, only one high-dimensional Gaussian integral needs to be calculated. The time-varying reliability of the independent standard normal space is simplified to the time-varying reliability of the large-scale morphing wing mechanism as a whole. The results show that: the relative error between iTRPD and Monte Carlo simulation (MCS) in analyzing the time-varying reliability of large-scale morphing wing is only -2.842%, better than that of the conventional method TRPD. The number of invocations of the function is 415, much smaller than that of the MCS (1×109 times). The number of invocations of high-dimensional Gaussian integrals is thirty-five times for the conventional time-varying reliability method, but only one time for iTRPD. It can be seen that iTRPD has high computational accuracy and computational efficiency for time-varying reliability analysis involving multimodal distributions.