The risperidone maintenance treatment in schizophrenia study was designed to identify the duration of maintenance treatment required with an initial therapeutic dose in contrast to reducing the dose over time. This study investigated extrapyramidal symptoms (EPSs) in different risperidone maintenance treatment paradigms over 1 year. Clinically stabilized patients with schizophrenia (n = 374) were randomized to a no-dose-reduction group and 4-week and 26-week reduction groups, in which the dose was gradually reduced by 50% over 8 weeks and maintained. Extrapyramidal symptoms were assessed at baseline and monthly for 6 months, followed by every 2 months. The Simpson-Angus Scale of Extrapyramidal Symptoms-Chinese version assessed EPS severity. Data were analyzed by a generalized linear mixed model (GLMM). The frequency of EPS at baseline was 23.2%, 20.0%, and 21.3% in the 4-week, 26-week, and no-dose-reduction groups, respectively. Risperidone dosage, positive symptoms, and disorganized thoughts at baseline predicted development of EPS. The GLMM indicated that a significant decrease in EPS was maintained, and different trajectories occurred over time across groups. In the 235 patients who continued treatment after 1 year, the incidence of EPS decreased to 4.1%, 2.8%, and 10.0% in the 4-week, 26-week, and no-dose-reduction groups, respectively, whereas the numbers of dropouts because of intolerable EPS were not significantly different (4.8%, 6.7%, and 6.2%, respectively). These novel findings indicate EPSs were tolerable and differentially decreased depending on the dose paradigm during the 1-year treatment period. Future studies should implement a GLMM to investigate antipsychotic adverse effects during long-term treatment.