In this paper, a new method is proposed to maximize a margin to the saddle node bifurcation in power system voltage instability. In recent years, the emergence of power markets with renewable energy brings about a lot of uncertainties to power system voltage instability. There is high possibility that voltage instability occurs for lack of nodal power injections. To make power system more secure, it is necessary to know how much current power system conditions are close to the bifurcation point and take control of power systems to escape from it appropriately. This paper proposes a new method for maximizing a margin to the bifurcation with capacitor banks of discrete variables. The problem formulation may be expressed as one of combinatorial optimization problems. After evaluating the bifurcation with the continuation power flow calculation, the optimal control of the capacitor banks is made by DEPSO (Discrete Evolutionary Particle Swarm Optimization) of evolutionary computation. To take account of uncertainties of power system conditions, MCS (Monte Carlo Simulation) is carried out for a power system with wind power generation unit. The proposed method is successfully applied to the modified IEEE 30-node system.