Synchronization analysis of electroencephalogram (EEG) and electromyogram (EMG) could reveal the functional corticomuscular coupling (FCMC) during the motor task in human. A novel method combining Gabor wavelet and transfer entropy (Gabor-TE) is proposed to quantitatively analyze the nonlinearly synchronous corticomuscular function coupling and direction characteristics under different steady-state force. Firstly, the Gabor wavelet transform method was used to acquire the local frequency-band signals of the EEG and EMG signals recorded from nine healthy controls simultaneously during performing grip task with four different steady-state forces. Secondly, the TE of local frequency-band was calculated and the unit area index of the transfer (ATE) was defined to quantitatively analyze the synchronous corticomuscular function coupling and direction characteristics under steady-state force. Lastly, the effect of EEG and EMG signal power spectrum on Gabor-TE analysis was explored. The results showed that the coupling strength in the beta band was stronger in EEG→EMG direction than in EMG→EEG direction, and the ATE values in the beta band in EEG→EMG direction decreased with the force increasing. It is also shown that the difference in TE values of gamma band present a varying regularity as the increase of force in both directions. In addition, EMG power spectrum was significantly correlated with the result of Gabor-TE inspecific frequency band. The results of our study confirmed that Gabor-TE can quantitatively describe the nonlinearly synchronous corticomuscular function coupling in both local frequency band and information transmission. The analysis of FCMC provides basic information for exploring the motor control and the evaluation of clinical rehabilitation.