Cases of COVID-19 in Latin America and the Caribbean were discovered in early March 2020. In Colombia, the National Institute of Health indicates that since the beginning of the pandemic, more than 3.8 million cases of people with symptoms of the COVID-19 virus have been reported. Departmental and municipal authorities have taken measures to reduce the infection rate, including Resolution 385 of 2020. The uncertainty is that once these new measures have been decreed, how do we know that they will really help to mitigate the number of infections and deaths registered per day? This motivated the design of a program developed in Python programming language, using the Linear Regression algorithm to predict the number of COVID-19 cases per day according to the data analyzed for Colombia. The algorithm yielded an accuracy between 90% and 99% in several tests performed, which allows to affirm that the algorithm satisfactorily predicts the COVID cases for the following day, obtaining an early forecast of the daily contagion, thus seeking to generate actions in the detection of patterns, which in a certain way are able to prevent or predict future events against the virus.