In this work a torque assessment approach is presented to estimate FES-induced torque no matter the muscle is fatigued or not in interrupted stimulation. This approach is based on evoked EMG (eEMG) signal and use peak-to-peak (PTP) amplitude and second phase area (SPA) to correlate eEMG and torque. The mapping function is built with pooled data of recruitment test and fatigue test in two complete paraplegic subjects. In this model, PTP is mainly responsible to the effect of recruitment level, SPA is mainly responsible to the muscle fatigue phenomenon. The model is fitted to the measured recruitment curves and measured fatigue curves and shows good fits, with R2 values above 0.95 and 0.93 for each subject. This suggests that it is possible to use processed eEMG signal to change stimulation intensity to obtain desired trajectory even when stimulated muscle is fatigued or to inform the fatigue condition.