Industry 4.0 promotes the use of Cyber-Physical Systems (CPS) to improve production efficiency through seamless data exchange between virtual and physical components. However, in manual labor-driven environments, discrepancies between virtual stock data and actual material usage can create challenges for accurate production monitoring. This study focuses on addressing these discrepancies by integrating a stock-taking method into a production monitoring system. The system was implemented in an air conditioning train car assembly workshop, where differences of 2–3% between the predicted virtual stock and real-world quantities were identified. By applying the stock-taking method, virtual data were recalibrated to reflect real-time stock levels more accurately. The system's ability to track material usage and losses allowed for significant improvements in inventory accuracy, with immediate updates provided to the CPS. This approach minimizes human error in manual operations, ensuring that material predictions are more aligned with actual consumption. The results show that the implementation of the stock-taking method reduced the margin of error in stock predictions, improving overall production decision-making. These findings suggest that this method can enhance stock accuracy in manufacturing sectors, particularly in developing countries where manual labor is predominant. This study provides practical implications for optimizing material management and reducing production costs by leveraging CPS integration with stock-taking methods.