The order-picking optimization of single-line fixed shelf similar to the symmetric and asymmetric traveling salesman problem is analyzed and proved. Then a new modified Lin-Kernighan (LK) algorithm is studied and applied it to the order-picking processes of the fixed shelves. The results of simulation indicate that the algorithm can reduce the computational complexity of time and space greatly, and also can find out the optimum solution quickly and steadily. The variation of the number of the order-picking points in wide range required by different tasks can be satisfied as well.