The aim of the nationwide retrospective matched cohort study was to evaluate health service utilization and medical costs between patients with schizophrenia who received long-acting injectable (LAI) risperidone and those who took risperidone orally. Data were sourced from the 2008 to 2013 Psychiatric Inpatient Medical Claim Dataset in Taiwan. The sample selection process was performed by propensity score matching. Finally, there were 691 patients in the exposed cohort and 1382 patients in the unexposed cohort. Each patient was individually followed for a 1-year period. Two-part models and generalized estimating equations were used to evaluate health service utilization and direct medical costs of patients. Analytical results showed that patients receiving LAI risperidone had used outpatient services significantly more, had greater hospital admissions, and had shorter lengths of stay than those who took risperidone orally. Furthermore, compared with their counterparts in the unexposed group, patients in the exposed group had incurred higher medical costs because of costs incurred from increased utilization of outpatient service and hospital admissions, under the special context of the healthcare system in Taiwan, a single-payer universal health coverage system with low copayment rates. In summary, this study suggested that patients with schizophrenia treated with LAI risperidone had shorter lengths of stay, higher medical costs largely because of increased utilization of outpatient service and hospital admissions, compared with those who took risperidone orally.