With the increasing penetration of renewable energy, the capacity allocation of microgrids needs to balance economics and robustness against uncertainty. Aiming at the problem of insufficient modeling of the impact of low-frequency fluctuation components on capacity planning in traditional methods, this paper proposes a low-frequency microgrid capacity allocation method based on Empirical Modal Decomposition (EMD) and two-stage robust optimization. First, the high-frequency and low-frequency components of renewable energy output and load demand are decoupled by EMD decomposition, and the deterministic trend of the low-frequency component is extracted; second, a two-stage robust optimization model is constructed: The first stage determines the base capacity of the equipment based on the low-frequency component, and the second stage utilizes KKT based on the low-frequency determined capacity to determine the minimum cost of capacity allocation, and finally, the algorithm is solved by the Columns and Constraints Generation (C&CG) iterative solving algorithm. Simulation results show that the proposed method reduces the allocation cost compared with the traditional robust optimization, ensures the economy under extreme conditions, and significantly improves the robustness of the system against power fluctuations.