In response to the high complexity and low accuracy of current facial expression recognition networks, this paper proposes an E-MobileNeXt network for facial expression recognition. E-MobileNeXt is built based on our proposed E-SandGlass block. In addition, we also improve the overall performance of the network through RepConv and SGE attention mechanisms. The experimental results show that the network model improves the expression recognition accuracy by 6.5% and 7.15% in RAF-DB and CK+ datasets, respectively, while the parameter and floating-point operations decreased by 0.79 M and 4.2 M compared with MobileNeXt.