The invention discloses an unmanned plane real time obstacle avoidance method based on the binocular visual sense technology; the method comprises the following steps: 1, using a binocular camera to gather parallax images, carrying out three dimensional reconstruction, and calculating the depth; 2, increasing the threshold, and only identifying obstacles in a scope smaller than the obstacle avoidance radius Zmax; 3, traversing the images, and calculating the obstacle condition of each pixel point; 4, calculating the obstacle filling rate in each grid unit; if the filing rate is >50%, an obstacle is considered to exist; if the filling rate is between 10-50%, an obstacle is unknown and risks exist; if the filling rate is smaller than 10%, the condition is considered safe; 5, analyzing the safety area size and distribution conditions, thus outputting front-back, left-right and up-down motion speeds, and a rotary angle rate of the controlled unmanned plane; 6, allowing a user to control aremote controller while carrying out binocular visual sense obstacle avoidance, thus forming extra front-back, left-right and up-down motion speeds v of the unmanned plane, forming the rotary angle rate [omega], naturally stably fusing with the flight control system, and fast and stably carrying out autonomous obstacle avoidance.