Abstract: The wireless service in complex environment shows the trend of data intensive development in recent years, which puts forward higher requirements for the data transmission capability and interference countermeasure capability of wireless communication systems. The existing spread spectrum and frequency hopping communication systems achieve interference countermeasure capability by sacrificing time-frequency resource utilization, but can not meet the demand of high-speed data transmission in complex environment. Therefore, this paper proposes a deep learning based narrowband interference suppression technology of non-spread-hopping spectrum communication system. On the basis of non-spread-hopping spectrum communication system, the frequency-domain notch filtering module and deep neural network module are cascaded at the receiver end to effectively suppress narrowband interference, improving data transmission rate and interference suppression ability at the same time. In particular, the deep neural network is used to reconstruct the expected signal from the distorted signal by the frequency-domain notch filtering module. The experimental results show that the proposed algorithm has a lower bit error rate than the traditional frequency-domain notch filtering algorithms, and the well-trained deep neural network can generalize to the scenarios with differences of signal power, interference power, interference frequency band, interference waveform, etc.