Abstract: A multi-constraint optimization model is built to minimize horizontal container transportation time, quay operation time and container dumping time in the yard under the condition of maintain the seaworthiness of the ship. The Improved Cuckoo Search algorithm is used to solve the problem. The nest information is extracted from the population as a whole, the contemporary optimal solution and its own individuals in the Levi flight stage and a variety of location updating modes are established. The discovery probability P_a is controlled by dynamic mechanism. The optimal solution is found by calculating the fitness function value of each net. Experiments are carried out with a typical data set. For the cited data, the optimization result is 25% higher than that produced by conventional cuckoo algorithm. Which proves the feasibility and effectiveness of the method.