The capacitor's health status is receiving increasing attention. However, the commonly used failure-of-physics modeling methods reveal the average performance of capacitor aging failure, but cannot explain the stochastic phenomena in practical applications. Therefore, this paper proposes a capacitor health status evaluation method considering stochasticity from multiple independent sources. Firstly, the stochastic phenomena and sources in the capacitor failure process are summarized comprehensively, and the characterization methods of internal parameter randomness and external condition randomness are given respectively. Then, Monte Carlo simulation and dynamic stress-strength interference theory are used to solve the modeling of aging failure and sudden failure, and the series model theory is used to form a composite failure probability to describe the health state of capacitors. Finally, the effectiveness of the proposed method is verified by accelerated aging experiments.