To address the issue of maximizing the underwater target detection efficiency of Unmanned Underwater Vehicle (UUV) formations, a research on the optimal design methodology for UUV formation under multiple constraints is conducted. Firstly, a grid-based UUV detection probability model is established to describe the search capabilities of UUVs. Then, considering constraints such as communication distance, communication angle and so on, a formation search performance evaluation model is designed for typical mission scenarios. By taking the UUV formation search performance evaluation model as the objective function and transforming the formation design issue into the problem of maximizing the search performance evaluation value, a novel hybrid optimization algorithm called Graded Incomprehensible but intelligible-in-time Logic Algorism (GILA) is proposed to solve for the optimal search formation. GILA is built upon the Incomprehensible but intelligible-in-time Logic Algorism (ILA) and its main improvements to the ILA are: employing good-point set to initialize the population; introducing dynamic weights to enhance the search ability of the algorithm; and integrating tabu search mechanism to prevent the algorithm from falling into local optima. Finally, the performance improvement effect of the GILA proposed in this paper is analyzed through the performance comparison simulation test of the optimization algorithm and the simulation test of UUV formation optimization design, and the correctness of the UUV multi-constraint formation design method based on GILA algorithm is verified.