To address the problem of anomaly detection and root cause tracing in cigarette production rate data, a novel machine learning algorithm was proposed. This study employs a variety of algorithms to detect anomalies in silk yield data, culminating in a final result derived through a weighted method. Following the identification of anomalous data, key factors are pinpointed using correlation and regression analysis algorithms. Experimental results demonstrate that this algorithm excels in identifying data anomalies and tracing their origins, significantly contributing to the enhancement of tobacco production yields. The application of machine learning in anomaly detection and root cause analysis within the tobacco industry represents a significant advancement in production efficiency and quality control. By accurately identifying anomalies and their underlying causes, this algorithm ensures higher precision in monitoring production processes and facilitates proactive adjustments to maintain optimal yield levels. The integration of multiple algorithms and a weighted method enhances the robustness and reliability of the anomaly detection process. Ultimately, this study provides a valuable tool for improving operational effectiveness and production outcomes in the tobacco industry, highlighting the broader potential of machine learning applications in industrial settings.