Antipsychotic treatment discontinuation is a major challenge in the treatment of first-episode schizophrenia (FES) patients. However, the rate and predictors remain unclear. Five hundred and sixty-nine FES patients were randomized to risperidone (n = 190), olanzapine (n = 185) or aripiprazole (n = 194) in a six-site study in China with 1-year follow-up. Patients failing the initially assigned antipsychotic were switched to one of the other 2 antipsychotics. By 52 weeks, 47.1% of FES patients discontinued all antipsychotics. In the 8-week acute phase, an antipsychotic switch was protective against antipsychotic discontinuation, whereas higher positive symptoms at the endpoint predicted discontinuation. In the maintenance phase, discontinuation was predicted by male gender and higher CGI-S score at the endpoint. The findings indicate that in China nearly half of patients with FES discontinued antipsychotic treatment during one year treatment. Clinicians should employ strategies other than medication choice to keep them from discontinuing.