Objective This study investigated the effects of metacognitive training (MCT) on improving psychiatric symptoms and cognitive function in patients with schizophrenia who were hospitalized for long periods. Methods The participants were long-term inpatients with schizophrenia, hospitalized in a private psychiatric hospital in Japan. Participants were randomly assigned to either the occupational therapy (OT)+MCT group or the OT-alone group. The OT+MCT group received 16 weekly MCT sessions, each lasting 60 minutes, over a period of four months. The Japanese versions of the Montreal Cognitive Assessment (MoCA-J), Positive and Negative Syndrome Scale (PANSS), Beck Cognitive Insight Scale (BCIS), and Global Assessment of Functioning (GAF) scores were compared before and after the intervention using a two-way repeated measures analysis of variance. To examine age-related effects, a correlation analysis was performed between participants' age and their MoCA-J total score. Furthermore, the OT+MCT group was stratified by median age (68 years), and changes in each outcome measure were compared between the groups. Results The 41 participants had a mean age of 69.22 years, ranging from 37 to 79 years, with 21 assigned to the OT+MCT group and 20 to the OT-alone group. There were no dropouts during the study period. The mean MoCA-J Total score at baseline was 17.67 (SD = 5.37) in the OT+MCT group and 14.85 (SD = 5.95) in the OT-alone group, indicating mild cognitive impairment in both groups. Four months after the intervention, both groups demonstrated an upward trend in MoCA-J scores. Notable improvements were seen in visuospatial/executive and total scores over time in both groups. However, no significant between-subject effects or interaction effects were detected. Additionally, a significant negative correlation was observed between age and the MoCA-J total score. In the PANSS, scores decreased in both groups at the post-intervention assessment, and significant differences were observed in general psychopathology and total score categories due to within-subjects factors. Negative symptom scores significantly differed between groups; however, no interaction with time was observed. Participants aged 68 or older showed greater reductions in PANSS scores. While no significant differences were observed in positive symptoms, significant improvements were observed in negative symptoms, general psychopathology, and overall PANSS scores. No significant differences were observed between the two groups in either the BCIS or GAF scores. Conclusions These findings suggest that MCT is feasible for long-term hospitalized elderly patients with schizophrenia and may improve psychiatric symptoms when combined with OT. Although OT+MCT did not significantly improve cognitive function, the results suggest it may be particularly beneficial for older patients in managing persistent psychiatric symptoms. Future studies should investigate the cognitive effects of MCT in larger samples, determine optimal treatment duration and frequency, and explore potential interactions with medications.