In this paper, we consider the frequency assignment problem (FAP) in a cellular mobile communication system under the assumption that there is no channel interference between two cells separated by more than a certain distance. This special structure is observed in most cellular systems. To handle the considered FAP, we use the pattern approach which fits naturally to the problem. Based on this approach, we are able to formulate the considered FAP into a manageable optimization problem and propose a two-phase heuristic algorithm for the problem. Computational experiments show that our algorithm performs much better in both solution quality and computational time than the recently developed algorithms for FAP. Since the considered FAP well reflects most cellular systems, our algorithm can be applied to many practical situations.