Objective: This study examined whether extrapyramidal symptoms predict incidence of tardive dyskinesia 1 year later. Method: Simple, global measures were used to rate extrapyramidal symptoms and tardive dyskinesia in a prospective, observational health outcomes study. Baseline and 3-, 6-, and 12-month data on 9,298 patients were analyzed by using a Cox proportional-hazard model. Onset of tardive dyskinesia was examined in two groups: 1) no tardive dyskinesia at baseline (broad risk set) and 2) no tardive dyskinesia at baseline and 3 months (narrow risk set). Results: Baseline extrapyramidal symptoms predicted later onset of tardive dyskinesia (broad risk set: hazard ratio=2.0, narrow risk set: hazard ratio=1.6). In analyses adjusted for age, gender, and medication exposure, this effect size was not reduced. About half of patients who developed tardive dyskinesia had earlier extrapyramidal symptoms. Conclusions: Although the association of tardive dyskinesia and extrapyramidal symptoms is significant, extrapyramidal symptoms do not robustly identify individuals at high risk for tardive dyskinesia. However, drug regimens and disease processes that increase extrapyramidal symptoms are likely to result in increased risk of tardive dyskinesia.