Introduction: Antipsychotics are widely prescribed for patients with schizophrenia. The Brazilian public health system provides these patients free of charge to patients and it is pertinent to evaluate their benefits. Objective: To evaluate the effectiveness of olanzapine and risperidone in the treatment of patients with schizophrenia in the real world and assessing risk factors for their discontinuation through a national non-concurrent cohort with 16 years of follow-up. Methods: Three SUS administrative databases were integrated by deterministic-probabilistic linkage. After, patients were matched (1:1) for psychiatric hospitalization, year of receiving the antipsychotic, sex and age, considering either olanzapine or risperidone at study entry. Kaplan-Meier was used to estimate the cumulative probabilities of discontinuation of treatment and associated factors were identified. Sensitivity analyzes were performed. Results: 3416 pairs of patients were included. Olanzapine had a longer time until discontinuation of treatment (p = 0.021), and risperidone had a higher risk of discontinuation (p = 0.021). Among patients persistent for at least 24 months, there was no statistically significant difference. Conclusion: Olanzapine demonstrated superior real-world effectiveness over risperidone, in terms of survival and psychiatric hospitalization. This superiority was not sustained in all analyzes.