Present track association algorithms in distributed multi-target passive tracking system always adopt theory in multi-sensor information fusion, in other words, only state informations of targets are concerned. A new track association algorithm is proposed based on information fusion of multi-feature such as frequency pulse-width and Pulse Repetition Interval(PRI) in distributed multi-target passive tracking system by applying Dempster-Shafer(D-S) evidence theory. The algorithm is of such advantages like faster association speed, higher correctness probability and it can be adapted to circumstance of dense targets due to the use of multi-feature information. The simulation proves that the proposed algorithm is superior to weighted track association.