Manual material handling remains a physically demanding activity with significant implications for workforce productivity and occupational health. While laboratory-based assessments exist, field-ready systems for real-time monitoring at the point of work remain limited. This study developed an edge-based predictive framework for estimating workforce productivity and physiological load without cloud dependency. The system integrates wearable sensors, including heart rate monitors, pulse oximeters, sphygmomanometers, and goniometers, with a locally deployed multivariate regression model designed for low computational overhead and minimal latency. Following ethical approval and informed participant consent, the framework was validated using 390 male participants aged 20–39 years. Participants performed lifting, lowering, and carrying tasks using wheelbarrows, carts, concrete blocks, and head pans under field and staircase conditions. Physiological and environmental variables were acquired at a sampling frequency of 1 Hz and processed through a GUI-embedded predictive algorithm. The system estimates energy expenditure (EE), maximal oxygen uptake (VO₂max), and task-specific work capacity. Experimental validation showed prediction errors of 7.39%, 8.54%, and 9.89% for VO₂max, energy expenditure, and workforce productivity, respectively, indicating strong agreement with measured values. The framework demonstrates the potential of edge-enabled wearable sensing systems for real-time ergonomic assessment, occupational risk reduction, and productivity optimization in manual labor environments.