The focus is on predicting the success of movies by using machine learning algorithms. The following systematic review performs the data collection process using the PRISMA methodology and the PICOC method employed to structure the search equation. The main objective addresses the analysis of the different research focused on effective algorithms for the prediction of box office success of movies using machine learning to help producers or investors develop the planning of their productions with a focus on profitability to avoid millionaire losses. The literature search results revealed a diverse set of algorithms: success classification, feature preprocessing, regression models, and models based on text, visual, and product features. Likewise, various algorithms were compiled that highlighted their effectiveness, such as Artificial Neural Networks (ANN), Act-direct, Support Vector Machine (SVM), Naive Bayes, XGBoost, and Random Forest. In conclusion, the RLS highlights the effectiveness of machine learning algorithms in predicting movie success, highlighting the need to consider specific factors such as budget, star value, and pre-release data. It also determines that the most efficient technique was XGBoost. It also highlights the lack of comprehensive research in this area, suggesting an opportunity to develop future research and methodological analysis, especially in crisis contexts, such as pandemics or strikes in the film industry.