In this paper, a multi-objective particle swarm optimization (MOPSO) with an initial non-inferior solution from maximizing directivities of conformal antenna array is proposed to get Pareto optimal solutions for desired multi-beam and low sidelobes. The proposed algorithm first uses multi-objective decomposition strategy to get a non-inferior solution by the weighted linear combination of multiple single-beam optimal solution. Then, the local search strategy based on the particle space and target space constraints at the same time in MOPSO is designed to achieve optimization of multi-beam and sidelobes. Results indicate that the approach can effectively get Pareto optimal solutions for multi-beam forming with low sidelobes of the satellite conformal array antenna.