An on-line waypoints re-planner was presented .It is capable of adjusting trajectory when some targets are detected. A mathematics model of multi-objective optimization under multi-constraint for re-planning was built. Advanced NSGA-II was adopted to find the Pareto non-dominant route set, aiming at maximizing weighted image quality of each target and minimizing route length under constraints. By using the general image quality equation(GIQE), image quality of each target was forecasted. Constraints were determined by field-of-regard(FOR) of sensor, dynamic motion of the vehicle, image quality demand, safety distance etc. Fuzzy selection was applied to select the optimal route from the Pareto non-dominant route set refer to current weights of weighted image quality, route length and risk given by expert system. The simulated result showed the rationality and intelligence of the method.