A novel spatial-time-frequency data association (STFDA) algorithm is proposed by taking advantage of the time-frequency separation between different frequency-hopping signals, which uses the time-frequency center as the basis of association. The time-frequency characteristics extracted from time-frequency analysis were used to design appropriate core function and construct a simplified TF-MUSIC algorithm. The Hungarian algorithm was incorporated in STFDA to search the best association scheme according to the cost functions from the time-frequency distance of different signals. Appropriate window function designed on time-frequency distance can prevent the side robe leakage from decreasing the precision of direction estimation, and improve the association correction probability. The simulation results confirm the feasibility and superiority of the proposed algorithm.