In view of the shortcomings of grasshopper optimization algorithm(GOA), such as insufficient exploration ability in the early stage and insufficient development ability in the later stage, this paper proposes a new adaptive chaotic GOA (ACGOA) by using curve adaptive adjustment strategy to replace the existing linear adaptive strategy, and using uniform chaos optimization to update the grasshopper position, so as to further improve the comprehensive ability of grasshopper swarms in the early exploration and later development, and better realize the balance between global search and local optimization ability. The effectiveness and stability of ACGOA in optimization calculation are tested by using a variety of single/multi peak benchmark function to get the extremum, and then applied to pattern synthesis of linear array for the first time, the target optimization performance of ACGOA is verified by suppressing the maximum SLL of linear array. The results show that compared with other existing algorithms, the algorithm has lower maximum sidelobe level and narrower first null beam width. The ACGOA results and convergence test show that the algorithm has superior applicability in antenna optimization.