Information and communication technology has revolutionized numerous sectors, including commerce, where prediction and data analysis play a crucial role in optimizing operations and maximizing sales. In this context, the article focuses on demonstrating the implementation of a web and mobile system to predict the sales of ornamental fish at the “GRANLEVEIN” farm, where various algorithms were compared in order to determine the most accurate prediction method. The methodology employed was Extreme Programming (XP), which aided in the planning phases and flexible identification of requirements to design the software. Regarding data science-oriented analysis, the Hefesto 2.0 methodology was used, and technologies such as JavaScript, HTML, PHP, Bootstrap, and the ml.js library were employed to build the Data Warehouse. The multiple linear regression model showed a 75% accuracy, standing out as the most suitable tool for this purpose compared to other algorithms like Naive Bayes and Random Forest. The significant variables included in the model were the type of fish, the selling price, promotions, and customer demographics, ensuring that the main influences on sales behavior were captured. Finally, it is concluded that the employed algorithm achieved greater accuracy concerning the farm’s input information, which helped improve the sales of ornamental fish for forecasting and business performance.