Objective. Rather than officially confirmed COVID-19 death counts, excess mortality is a more comprehensive measure of the full impact of the COVID-19 pandemic. This study determines an estimate to the real number of COVID-19 deaths in Mexico in the period from 2020 to 2023.n-hospital delivery-related maternal mortality and sociodemographic factors influencing this mortality in Ecuador during 2015 to 2022. Methods. Using historical mortality data 2011 to 2019 provided by INEGI (National Institute of Statistic and Geography), we made a machine learning prediction for the number of deaths by any cause at years 2020 and 2021 in Mexico. Results. The differences between INEGI officially reported numbers of total deaths by any cause for these years minus the machine learning predictions for deaths by any causes (trained data 2011-2019 before COVID-19) are compared to the officially reported COVID-19 deaths, thus scaling factors are obtained. The estimate for the real number of COVID-19 deaths in Mexico is accomplished by multiplying the scaling factor (mean of two values) times projected values for the number of officially reported COVID-19 deaths at the end of July 2023. Conclusions. Machine learning algorithms have been employed to obtain first projections for the number of deaths in Mexico by any cause for years 2020 and 2021, and then a projection for the number of officially reported COVID-19 cases at the end of week 30 2023. Thus, we calculate an estimate of 760 398 for the real number of COVID-19 deaths in Mexico at the end of week 30 of 2023 (July 29th, 2023).