The objective of this study is to design a data-driven model for extracting effective rules influencing the product acceptance rate in food industry production lines. To this end, first, 25 variables related to lean manufacturing principles were identified, and using the fuzzy Delphi method, six key variables were selected, including the percentage of conformity of raw materials with quality standards, production scrap rate, exceptional approval rate of the final product, average raw material storage time, work shift, and product return rate. Subsequently, a database consisting of 4,200 real records from two years of production line performance was collected, and after data preprocessing, modeling was conducted using machine learning algorithms. Among these, the decision tree algorithm (CART) was employed as the primary model for extracting interpretable decision rules, and its performance was compared with other classification algorithms. The results indicated that the “percentage of raw material conformity” and the “scrap rate” had the greatest impact on product acceptance. The decision tree model achieved an accuracy of 91% in classifying samples into acceptance and non-acceptance categories. Moreover, the extracted rules revealed transparent and interpretable relationships among the variables and provided a basis for developing decision-support patterns for production and quality control managers. Overall, the findings indicate the effectiveness of lean production–based data-driven approaches in improving quality prediction and reducing waste in the case study of this research.