This study proposes a novel approach to monitor the fatigue levels of construction rebar benders by measuring chemical biomarkers using sweat sensors. Fatigue resulting from dehydration and energy depletion can severely endanger the safety and health of construction workers. Sodium, lactate, glucose, and sweat rate were chosen as detectable biomarkers in this study, as their concentrations can indicate hydration status, energy consumption, and electrolyte balance, making them suitable for fatigue monitoring. The results were used to construct a fatigue model using supervised machine learning approaches. Construction rebar experiments were conducted while the sweat-based biosensors were applied to rebar workers to evaluate their fatigue with five different classifiers, demonstrating accuracy rates ranging from 71.43% to 96.43%. The results suggested that sweat-based biomarkers offer a noninvasive and accessible fatigue monitoring alternative. This can potentially help alleviate fatigue-related adverse ill effects like dehydration or cramping by enabling instant fluid or nutrient supply recommendations during construction manual tasks. It also provides valuable insights into the physiological effects of rebar work. Besides, this study presents a valuable model for predicting workers’ fatigue levels, which could be applied in the construction industry to improve workers’ safety and productivity. Furthermore, the study highlights the importance of maintaining appropriate hydration, nutrition, and electrolyte balance during physically demanding tasks like construction manual work.