In this study conducted as part of the project “Evaluation of Predictive Production Models Using Curves of the Main Diseases in National Cacao and Native Mango Crops in Ecuador”, we describe the development and integration process of an online platform based on machine learning, aimed at calculating the productivity of cacao crops in Ecuador. To generate precise predictions of future cacao yields, the system employs historical production data and relevant climatic variables such as temperature, precipitation, and humidity. Various regression models were utilized, among which Random Forest stood out for its ability to handle multiple variables and generate accurate estimates, achieving a coefficient of determination (R2) greater than 85%. The developed platform comprises a mobile application for on-site data collection and a web application that empowers agricultural managers to visualize yield projections and make strategic decisions using this data. The use of open-source technologies like Python, Django Rest Framework, and React.js has contributed to accelerating the creation of an accessible and efficiently scalable solution. The study’s results demonstrated that the platform has the capability to accurately predict cacao productivity up to five months in advance, which contributes to optimizing decision-making processes. We have identified a positive relationship between the level of Monilla infection and harvest yield, providing relevant data to optimize plantation management.