This study proposes an indirect approach that estimates roll force from motor current signals in a rolling mill. Motor current is first converted to motor torque using an induction-motor equivalent-circuit model, then to roll torque via the gear ratio, and finally to roll force through a torque-arm relationship. A laboratory-scale rolling mill was designed and fabricated to experimentally validate the approach. Two torque-conversion schemes were examined: Method A, which determines the slip of the induction motor from measured rpm and recalculated motor parameters, and Method B, which estimates slip from measured motor current and applies a finite element (FE)–based response surface function to calibrate the converted torque. The converted roll torques were validated against FE analysis, and the resulting roll forces were compared with load cell measurements under various rolling conditions. Deviation, defined as the average difference between the FE-predicted torque and the converted torques, ranged from −11.9% to 28.8% for Method A and −7.2% to 13.8% for Method B. Roll force deviations from measurements ranged from −14.1% to 14.9% for Method A and −3.7% to 14.2% for Method B. Method A provided a straightforward and computationally light conversion route but was more sensitive to rpm-measurement noise, whereas Method B yielded smoother temporal behavior at the cost of FE-based calibration. The results demonstrate that both methods can reproduce the overall evolution of roll torque and roll force using only motor-side measurements, offering a practical foundation for real-time monitoring in rolling mills where stand-by-stand load cells are unavailable.