Virtual power plants (VPP) can provide considerable reserve capacity for power grid operation. Accurately evaluating and quantifying the reserve capacity of VPP is key to their participating in power grid regulation. However, the strong uncertainties associated with distributed renewable energy output, load consumption, electricity price, and other factors can lead to unreliable results when using traditional deterministic methods for reserve evaluation. To address this, a definition and evaluation method for reserve credit under multiple uncertainties is proposed based on the characteristics of VPP. First, the framework of VPP and the definition of reserve credit are introduced. Then, a VPP reserve provision model is constructed considering various resource aggregation. By applying Monte Carlo to model multiple uncertainties and using kernel density estimation, a set of reserve credit with different confidence levels is derived, effectively quantifying the reserve credit of VPP. Finally, a case study on a typical VPP is conducted. It demonstrates that the proposed method can effectively reflect the probability characteristics of the reserves that VPP can provide considering multiple uncertainties, providing dispatch agencies with more comprehensive and reliable reserve information.