The paper introduces an approach to designing and modeling an Internet of Things (IoT)-based machine learning control system for smart manufacturing factories in response to demands for optimization, flexibility, and eco-effectiveness. The system incorporates IoT for capturing actual process data in real time, ML for information processing and analysis, and control strategies for performing an optimal control over the underlying processes. For performance evaluation of the system under dynamic operation, simulation framework was designed implemented in MATLAB. Other quantifiable criteria, including control accuracy, response time, and energy used within the optimal control interval, were examined. The system proved its capability for handling steady flow of data, fine-tuning of control actions, and substantial minimization of energy demands. The findings provide evidence for the ability of IoT and machine learning methodologies in successfully integrating intelligent manufacturing. This study outlines how such systems exhibit great promise as the basis for future smart factories, which is supported by the roadmap established in this work to deepen and expand upon future research avenues such as real-world implementation, more effective techniques to negate noise interference, and high-level control strategies to improve scalability and robustness of such systems.