The importance of monitoring and quality prediction in the industrial sector has become essential to ensure the quality of products before producing them and reduce the scrape rate. The industries are looking for advanced tools and means to predict and optimize product quality. This paper proposes a novel model for quality prediction tasks using reinforcement learning (RL), and then compares its performance with machine learning (ML) algorithms studied in our latest article. The model was validated using real data from an industrial site in Morocco. The results show that ML algorithms, especially Multi-Layer Perceptron (MLP) achieve good accuracy, recall, and F1-score, demonstrating their effectiveness in predicting the quality of products, while RL shows solid performance but with some variability. Despite that, RL has a big potential in decision-making, it requires further optimization to match the performance of ML models. This highlights the strengths of ML in quality prediction tasks and the need for further improvement of RL techniques.