The intermittency in weather complicates the task of determining the optimal setting of protective relays in hybrid photovoltaic (PV)-wind-based microgrids. Classical overcurrent relays quite often maloperate during fluctuations in wind speed and solar irradiance. Motivated by the significance of reducing the sensitivity of microgrid protection against weather intermittency, this article proposes a protection scheme based on the combined framework of stochastic modeling of weather intermittency using probability distribution function and ensemble-based classifier. Considering the possible similar weather scenario to which both the distributed energy resources are subjected to, the covariability has been considered to develop the joint probabilistic model of variation in wind speed and solar irradiance. The model derived using the historical data allows incorporating the weather intermittency in the formulation of the protection algorithm. The data generated under intermittency are processed using a wavelet transform to derive discriminatory attributes. Using the derived attributes, a rotation forest-based classifier has been developed to perform fault detection/classification and faulty section identification. The scheme has been validated for a wide range of weather and fault scenarios. The comparison of the proposed scheme with other reported techniques reflects its significance in providing resilient and robust protection to the microgrid against weather intermittency.