Abstract: Maneuverability is one of the most crucial navigation characteristics of Unmanned Underwater Vehicle （UUV）， which directly impacts the safety and stability of the vehicle. Therefore， the evaluation and optimization of its maneuverability are key steps during the UUV design process. In this paper， a comprehensive mathematical model for maneuverability optimization of UUV is proposed， based on the AUTOSUB type UUV structure and adequate analysis of UUV maneuvering performance （mainly vertical and horizontal motion）. The objective function is constructed using the fuzzy evaluation method， with the weight of each subsystem function determined by the expert consultation method； the membership functions are obtained through statistical analysis of all the indicators and functions chosen in the fuzzy library； the Genetic Algorithm （GA） is incorporated with Tabu Search （TS） and forms the GA-TS algorithm， which can be further categorized into parallel and layered strategies called Parallel Genetic Algorithm（PGA） and Layered Genetic Algorithm （LGA）. Finally， the UUV maneuverability optimization program is written via C++ by integrating the mathematical model with three obtained algorithms. In contrast with PGA and LGA， the GA-TS algorithm shows high optimizing capacity. Also， the weight influence of each performance indicator on the system objective function is analyzed for horizontal and vertical motion subsystems， and the optimal weight distribution is presented.