Abstract: To solve the problem of rational location of repairing units and improve the timeliness of the battlefield-repairing, a new location model with double coverage standards was constructed based on Laplace criterion of uncertainty decision and aiming at the maximum of edge covering ratio of the whole road network by supposing that task demands occur at any point of the network. An improved genetic-tabu algorithm was designed to solve this model. A group of initial solutions was obtained by the area-dividing clustering method. To avoid the generation of too many infeasible solutions in searching, the heuristic strategy was adopted at cross choosing. The results show that the proposed algorithm has a low computational complexity, and the precision problem of the location model based on edge demands is solved without increasing the network dimensions.