The surging COVID-19 infections and deaths in some parts of the world like India and Brazil is an issue of concern in the COVID-19 response (Shet et al, 2020; Sun et al, 2020). New strains of the corona virus, SARS-COV2 have been seen to be more transmissible. The most worrisome issue is the death of young and economically productive people due to the new variants in countries like Brazil. Several previous studies have shown that the significant predictors of COVID-19 mortality are old age, hypertension and diabetes (Jordan et al, 2020; Wu et al, 2020; Becchetti &Salustvi, 2020; Ma et al, 2020; Hasan & Haque, 2020). Ghana reported the first 2 cases of COVID-19 on the 12^th March 2020. By 15^th of April 2020 the country had reported 641 positive cases of COVID-19 (WHO, 2020). The government managed the situation by imposing lockdowns, quarantine, isolation, contact tracing, hygiene practices, social distancing and treatment of COVID-19 cases (Ghana emergency, 2020; Zurek, 2020). In this study we propose to apply a machine learning algorithm to predict daily COVID-19 deaths in Ghana. The findings of this piece of work are envisioned to reveal the future trends of COVID-19 mortality in the country and stimulate an appropriate early response to the epidemic.