In this study, an algorithm is presented with the goal of enhancing the evaluation capacity of regression models in scenarios where predicting multiple output variables simultaneously is required. The proposed algorithm builds upon the extension of the similarity quality measure of a decision system, offering a novel and practical approach to address the unique challenges associated with Multi-Target Prediction (MTP) problems. The primary objective is to improve the accuracy and reliability of machine learning models in these specific contexts, ensuring that the algorithm can effectively capture the relationships between multiple dependent variables. This contributes to better model evaluation and improved performance in real-world applications where multi-output prediction is critical, such as environmental modeling, healthcare, finance, and multi-agent systems. Furthermore, the algorithm’s simplicity, adaptability, and cost efficiency make it suitable for integration into various machine learning pipelines, paving the way for future advancements in regression analysis, decision-support systems, and intelligent automation frameworks.