As one of the key issues of the automatic navigation of air vehicles, route planning isan important research field of robotics. It has been attracting a great attention boththeoretically and practically. According to various mission scenarios of air vehicles, thisdissertation addressed the following route-planning problems: (1) Modeling of routeplanning;(2) Off-line route planning;(3) On-line route planning;(4) Route planning basedon evolutionary;(5) Multiple paths planning;and (6) Cooperative route planning formultiple vehicles.In the modeling of route planning, three works are conducted: Firstly, a newrepresentation method of planning environment is proposed. By representing differentenvironment attributes in different data formats, not only a time-consuming reconstructionof search map is avoided, but also the environment attributes can be updated in real-timewhen the new information of the environment is available. Secondly, our flight route isconsisted of a series of 3-D waypoints. Thereby, an arbitrary precision can be obtained byadjusting the number of waypoints. Finally, a cost function of flight route that maximizesthe survival probability of the vehicles and reduces the computational burden is estableshed.In addition, a determination method of route weighting factors is also addressed based onfuzzy inference.In the research of off-line route planning, a new 3D route planner based on SAS(Sparse A* Search) is presented. By incorporating route constrains into search algorithm,this approach can efficiently prune the search space, shorten the search time, and generatean optimal solution. As a dynamic version of SAS, the DSAS (Dynamic Sparse A* Search)algorithm can be used for on-line route re-planning in flying environment with unknownthreat, and keep the optimality of the route at any time.There are two parts in the research of on-line route planning. For the vehicles towardsstill target, a new on-line real-time route planner is proposed. By on-line planning during theflight,this planner can dynamically adjust the time of search and make it satisfy therequirements of in-flight applications. Correspondingly, for the vehicles towards a movingtarget, a moving target route search algorithm is given, which reacts to every movement ofthe target. The convergence of the two planning algorithms is also discussed.Based on evolutionary computation, a new 3D route-planning algorithm for air vehiclesis presented. Combining the concepts of evolutionary computation with problem-specificrepresentation of candidates and genetic operators, the routes are generated in real-time andare able to take into account different kinds of mission constraints. Besides, it can determinethe number of the route nodes according to the environment automatically.In addition, the multi-route planning problem of air vehicles is discussed with theproposal of a new planner. By using the K-mean clustering algorithm, the number of resultantroutes can be inputted according to the environment and the requirement of the mission. Byclustering the routes according to their distribution in the space, the algorithm can generatemultiple separated routes. By using the technique of multi-population evolution, it is suitablefor parallel computation, which can speed up the computation process.At last, the cooperative route-planning problem of multiple air vehicles is addressed anda novel co-evolutionary coordinative route planner is developed. In this new planner, thepotential routes of each vehicle form their own sub-population, and evolve only within theirown sub-population, while the interaction among all sub-problems is reflected by thedefinition of fitness function. The Algorithm can take into account different kinds ofenvironmental information, handle various mission constraints, and generate desired 3Droute in real-time for each vehicle.