In order to reduce the false alarm rate of autonomous obstacle avoidance for intelligent forklift trucks in dried fruit warehouse and improve the utilization rate of warehouse space, combining with the characteristics of the operating environment of forklift truck, and basing on the laser sensor SICK-LMS111 and SICK-NAV350, a dynamic identification zone that the detection range was changed along with the speed and equivalent steering angle was set up to effectively detect obstacles in the direction of travel. Then, feature control point obtained by obstacles was taken as the segmentation point, and the quartic fiveorder quasiuniform Bspline curves were used to generate the obstacle avoidance path by combining front and rear segments of curve, which met the multiconstraint conditions such as the minimum turning radius, the continuous curvature and the maximum angular velocity of steering wheel. By carrying out the obstacle avoidance path planning tests on the straight road section and the turning in the warehouse, the results showed that the obstacle avoidance paths satisfied various constraints, the curvature was not more than 1.06×10-3mm-1, equivalent steering angle was not more than 60°, angular velocity of equivalent steering wheel was not greater than 1.05rad/s, which effectively verified the feasibility of the algorithm. It can be used as a reference for obstacle avoidance technology of similar large warehouse intelligent vehicles.