A sidelobe control method is proposed for robust adaptive beampattern based on worst-case performance optimization. Multiple quadratic inequality constrains are added to arbitrary sidelobe regions, which can achieve lower sidelobe level as well as adaptively suppress interference. The optimization model can be converted into convex second-order cone(SOC) programming issue, which can be solved efficiently using the existing interior point method. Simulation results demonstrate that in the case of beam pointing error, the approach proposed can obtain the prescribed low sidelobe without performance loss.