Abstract: Track-to-track association is the precondition of multisensor bias estimation and track-to-track fusion. In terms of global nearest neighbor the misassociation condition is proposed, and the in°uence of the null association cost, sensor biases, false alarms and missed detections on the generation of misassociations is discussed. Analysis and simulation demonstrate that the impact of misassociations on multisensor bias esimation is quite di®erent from that of random noises, and misassociation could bring large estimation bias.