To solve the problem that artificial bee colony algorithm is good at exploration and neglect exploitation, this paper proposes an improved artificial bee colony algorithm based on genetic model, which has been successfully applied to array synthesis. Firstly, the global optimal solution is introduced into the neighborhood search process to guide the bees to find the best nectar source thus to accelerate the convergence of the algorithm. Secondly, to avoid the local optimization of the algorithm, the exploitation ability of artificial bee colony algorithm must be improved. The evolutionary mechanism of genetic algorithm is used for reference, and a genetic model is established to carry out genetic operation on the honey source after adopting the optimal retention, to enrich the diversity of honey source. The improved artificial bee colony algorithm is tested on a set of widely used numerical functions, and the experimental data show that the proposed algorithm has strong competitiveness compared with other algorithms. Then, the algorithm is applied to the sparse optimization of the linear array to reduce the peak sidelobe level of the array. The optimization is compared with other algorithms under the same array constraints. The simulation results further prove the effectiveness of the algorithm.