Surface Electromyogram (sEMG) is an indicator of fatigue progression during isometric or dynamic contraction of skeletal muscle. Estimation of fatigue index provides clinically relevant data for the diagnosis of neuromuscular disorders. The major challenge is that the signal is highly nonstationary upon dynamic contraction of muscles. Therefore, an advanced signal processing method is essential for the analysis of such signals to measure fatigue indices. Cyclo-nonstationary (CNS) analysis reveals the hidden periodicities of a highly nonstationary signal which is not firmly established for muscle fatigue analysis. In this work, an attempt has been made for the cyclo-nonstationary analysis of sEMG signal in biceps brachii muscle using Order-Frequency Spectral Correlation function (OFSC) method. For this, signals are recorded from fifty healthy volunteers with well-defined protocol. The preprocessed signals are equally partitioned into three segments namely, nonfatigue, progression of fatigue and fatigue. Further, OFSC is computed using the Welch-based estimator. In addition, the OFSC statistical features such as area under the curve, skewness, kurtosis and six entropies are estimated to evaluate fatigue condition with CNS analysis. The preliminary results show that OFSC features are able to differentiate the fatigue condition. The obtained results are statistically significant with p < 0.002. Therefore, OFSC-based CNS analysis can be used for the fatigue index estimation to diagnose neuromuscular disorders. © 2019 IEEE.