The uncertainties are depicted from aspects of time-sequence and randomness, so as to construct a model describing the photovoltaic(PV) power output and load with uncertainties. The derivation and application of locational marginal capacity cost(LMCC) has been improved, and the calculation method of LMCC with and without PV has been clarified. The sequential Monte Carlo simulation method is used to deal with uncertainties, and the calculation of LMCC is applied to guiding the sitting and sizing of distributed PV generation, based on which the PV planning model is established, and the corresponding algorithm flow is proposed. The model is applied in simple feeder analysis as well as IEEE33 node system planning to verify the effectiveness of the method proposed. Besides, the operational strategy of the grid company for PV is also preliminarily discussed, which indicates the direction for subsequent research.