The main objective of this research is to "improve production performance" in the spinning and single-ply yarn production section in the textile industry using artificial intelligence and appropriate machine learning algorithms and to apply an optimal algorithm to make the Business Intelligence in this industry. The data was analyzed by regression algorithms to improve the performance of yarn production with different fineness in the Worsted system and minimize stops, waste, and thus increase efficiency. The results show that machine learning models can successfully predict production performance improvement criteria in the textile industry to reduce waste and unplanned stops, thereby improving and increasing production efficiency in the single-ply yarn production stage in ring spinning machines in Guilan Spinning Company. In this study, the “Categorical Boosting Regression” model with R² Score and RMSE fitness indices of 0.938 and 22.182 respectively has the best accuracy and performance for data analysis and prediction. The advantage of using this model over other models is the reduced probability of overfitting, the ability to process batch data (classified) without the need for extensive preprocessing, high accuracy, fast processing speed, and less time for training. Accordingly, using the SHAP technique, it was determined that the most important factors affecting "production performance" based on their value in production performance are, respectively, yarn count, main cause of stoppage, amount of waste, duration of stops, and explanation of the main cause of stops. The main focus for “improving production performance” should be on “root causes of stop time” and “waste management” under production pressure conditions.