Recently, interest in forecasting events in various social fields has increased. Obviously, to make a realistic forecast it is necessary to have a specialized algorithm that takes into account many influential factors. This article proposes a weighted algorithm for predicting the results of football matches, according to which weighting values are assigned to certain factors to numerically evaluate the expected outcome of the match. The main criteria for predicting certain values are the status of the team (home or away), as well as the match age of the match relative to the predicted one (the earlier the match took place, the lower the weight coefficient is assigned to it). In the article, the main criteria for predicting certain values are the status of the team (home or away), as well as the prescription of the match (the earlier the match took place, the greater the weight coefficient) in relation to the predicted one. For the analysis of the value, the last fifteen matches of the teams participating in the predicted match are taken. If the current status of the team is home, then matches in which this team was home have a weighted coefficient of 0,7, and guest ones - 0,3. Similar values apply to the guest team (past guest matches have weight coefficient of 0,7, while home matches have weight of 0,3). The prescription of matches is divided into three categories: recent (weighting coefficient is 0,5), medium-old matches (weighting coefficient - 0,3) and late matches (weighting coefficient - 0,2). The work of this algorithm on a specific match is shown. As the analyzed value, the number of goals the teams can score is chosen. An analysis of the possible number of goals scored by each team was carried out by assigning weights to the current status of the teams and the prescription of the analyzed matches. Received data correspond to reality. To estimate the accuracy of the proposed method, an analysis of thousands of past matches was carried out. Conclusions are drawn about the correct work of the described method. An algorithm is considered correct if its accuracy is at least 80%. Based on the obtained accuracy value (81,2 % ≥ 80%), it can be concluded that the proposed forecasting method is correct.