Sending-end multi-terminal high-voltage direct current (MT-HVDC) systems are well-suited for large-scale renewable energy collection and transmission. However, the capacity planning for converter stations (CSs), which is directly correlated with their ability to convert renewable energy, remains a critical issue. In this paper, an optimal capacity planning method for CSs is proposed to maximize the converted energy (CE). The proposed method considers the uncertainties of photovoltaic (PV) generation and derives analytical formulas for stochastic CEs. The equal incremental rate (EIR) principle is employed to calculate the optimal capacity planning scheme, and then a general guideline for the capacity planning in stochastic scenarios is presented. Case studies are conducted to validate the effectiveness of the proposed method and the proposed guideline. The results demonstrate that the proposed method converts more renewable energy than the deterministic method.