The perovskite solar cell (PSC), as a game-changing breakthrough within next-generation electrical-optic devices, has become one of the most rapidly emerging market contenders. The free vibration analysis of the PSC is one of the most fundamental and crucial issues for the whole dynamic system, whereby undesirable vibrations can be attenuated to the lowest extent. In addition, manifold physically inspired uncertainties inherently exist in realistic engineering applications. Consequently, free vibration investigation based on the deterministic model without intrinsic randomness is merely an approximation of the actual structure behavior. Hence, this research presents the first work in implementing stochastic free vibration analysis of the PSC with multi-dimensional uncertainties possessing disparate distributions. An adaptive virtual modeling technique, namely the sequential design of experiment enhanced adaptive Kriging, is developed to depict the underpinned sophisticated relationship between physically inherent input uncertainties and concerned dynamic performance, and to offer adequate statistical information of structural response, including probability density function (PDF) and cumulative distribution function (CDF). Besides, the adaptive virtual model is competent to furnish accurate dynamic response forecasts for the PSC against rapidly varying conditions with greatly reduced computational costs, enabling the continuous information update and real-time dynamic performance surveillance system in modern engineering practice.