A method is applied to a server and for providing a link priority between the server and a storage device. The server calculates an average latency of each of a plurality of links between the server and the storage device; then determines a latency interval to which the average latency of each link belongs; and determines a priority of each link based on a priority of the latency interval to which the average latency of each link belongs. At least two consecutive latency intervals are preset in the server, each latency interval represents a range of an average latency, and a priority is set for each latency interval. For latency intervals of adjacent priorities, a minimum latency value of a latency interval of a lower priority is greater than a maximum latency value of a latency interval of a higher priority.