This paper proposes an improved computational algorithm for structure topology optimization. It integrates the merits of Evolutionary Structure Optimization and Level Set Method (LSM) for structure topology optimization. Traditional LSM algorithm has some drawbacks, for instance, its optimal topology configuration is largely dependent on the structural topology initialization. Additionally, new holes cannot be evolved within the updated topology during the optimization iteration. The method proposed in this paper combines the merits of ESO techniques with the LSM scheme, allowing new holes to be automatically inserted in regions with low deformation energy at prescribed iterations of the optimization. The nodal neighboring region is a good selection. For complex structures in which holes cannot be properly inserted in advance, the proposed method considerably improves the ability of LSM to search the optimal topology. In addition to achieving more accurate results, the proposed method also yields higher efficiency during optimization. Benchmark problems are presented to show the effectiveness and robustness of the new proposed algorithm.