The problem of goal-oriented obstacle avoidance for mobile robot in unknown environment is studied. By imitating the preview navigation behavior of human, the paper proposed a novel real-time fuzzy navigation algorithm for mobile robot., Firstly, the preview path was estimated based on the current scenario. The objective orientation of robot was calculated by synthesizing preview path tracking and obstacle avoidance. The angular velocity fuzzy controller was designed to tracking the objective orientation rapidly and accurately. The linear velocity fuzzy controller was designed for obstacle avoidance and high navigation performance by using the danger of collision, preview path curvature and angular velocity controllerpsilas output as input. The simulation experiment results show the validity and effectiveness of the navigator.