According to the recently proposed concept of PV-STATCOM that utilizes photovoltaic (PV) inverter as STATCOM, PV solar farm plays an increasingly important role in providing fast reactive power compensation, especially during the voltage recovery processes under voltage sag or post-fault conditions. To relieve the enormous economic losses caused by the inefficiency in voltage recovery scenario, the reactive power fast response characteristic of PV should be adequately considered in optimal siting and sizing of distributed generation. In this paper, a weighted voltage support ability index (WVSAI) based on sensitivity analysis has been firstly used to quantify the influence of PV-STATCOM on voltage recovery efficiency under voltage sag or post-fault conditions. With the WVSAI added to traditional distributed generator (DG) planning constraints, a modified optimal siting and sizing model is proposed for the targets of minimizing the total annual cost associated with the DGs to be planned. The proposed model is classified into the type of mixed integer nonlinear programming (MINLP) and then relaxed by second order cone programming (SOCP). After relaxation, commercial solver CPLEX through YALMIP platform is used to solve the relaxed model. The proposed model has been applied in IEEE 33-bus distribution system under scenarios with different WVSAI restrictions. Based on the optimal solutions of DGs' siting and sizing, PSCAD simulation tool is used to emulate voltage recovery processes of a specific sensitive load under voltage sag, and thus demonstrates the feasibility and effectiveness of the proposed model.