This paper addressed obstacle avoidance problem of a quadrotor in outdoor environment. Path planning was finished by employing classical Dijkstra algorithm and the controller of the quadrotor adopted integral backstepping method. In order to get a smooth trajectory with time optimal and acceleration constraints properties, quantic polynomials were utilized to generate trajectory file. The scheme is particularly practical and useful because of its small computation burden and feasibility. Simulations have been done to verify the performance of the whole system. As a result, the quadrotor successfully tracked the path without any collision with obstacles except that there are some small error in the final part of position tracking performance (±1m) and some small delay in the attitude tracking about 0.3 ms. In conclusion, the proposed method showed acceptable performance for quadrotor obstacle avoidance.