Objective: To investigate underreporting of immunization errors based on vaccination records from children under five years of age. Method: An epidemiological, cross-sectional analytical study, carried out through a household survey with 453 children aged 6 months to 4 years in three municipalities in Minas Gerais in 2021. A descriptive analysis was carried out, and the prevalence of the error was calculated per 100 thousand doses applied between 2016 and 2021. The magnitude was estimated of the association between variables by prevalence and 95% Confidence Intervals (95%CI). To analyze underreporting, State reporting records were used. Results: A prevalence of immunization errors was found to be 41.9/100,000 doses applied (95%CI:32.2 – 51.6). The highest prevalence occurred between 2020 (50.0/100,000 doses applied) and 2021 (78.6/100,000 doses applied). The most frequent error was an inadequate interval between vaccines (47.2%) associated with adsorbed diphtheria, tetanus and pertussis (DTP) vaccine (13.7/100,000) administration. Vaccination delay was related to immunization errors (7.55 95% CI:2.30 – 24.80), and the errors found were underreported. Conclusion: The high prevalence of underreported errors points to a worrying scenario, highlighting the importance of preventive measures.