The chikungunya virus (CHIKV) was first detected in Brazil in 2014 and has since caused major epidemics. In 2023, a major chikungunya epidemic occurred in Minas Gerais, one of the most populous states in the country. We analyzed the North and Northeast Health Macroregions of Minas Gerais, with 2.5 million inhabitants. The study used epidemiological data from the Ministry of Health’s laboratory surveillance, disease notification and mortality systems. A robust Poisson regression model was constructed to estimate expected monthly mortality. Excess deaths were calculated by comparing observed deaths with model estimates during the epidemic period. During the epidemic, there were 890 excess deaths attributed to chikungunya, translating into a mortality rate of 35.1/100,000 inhabitants. The excess mortality rate was significantly 60 times higher than the deaths reported by surveillance, with only 15 confirmed deaths. The correlation between excess deaths and laboratory-confirmed chikungunya cases was strong, while the correlation with dengue and COVID-19 was not statistically significant. The results highlighted the serious underestimation of chikungunya mortality by epidemiological surveillance. During the same year, only 420 chikungunya deaths were reported by all PAHO member countries. Routine epidemiological surveillance systems cannot capture the full impact of this disease. Excess mortality is a key measure for understanding the impact of epidemics, as demonstrated by COVID-19 and influenza pandemics. The study highlights the need for complementary tools to traditional surveillance to better assess impacts on morbidity and mortality and support priority setting in public health interventions.