In this paper, we present a deep neural network (DNN) based algorithm to restore the high-frequency (HF) component of the enhanced speech processed by Kalman filtering, where the DNN is applied for estimating magnitude of HF component from that of low-frequency (LF) component. The estimated HF component is obtained with the estimated magnitude given by DNN and the phase of the denoised speech. By incorporating our restoration algorithm into Kalman filter based method, our system is able to recover the HF component with better perceptual quality and less distortion. Experimental results demonstrate that our proposed system outperforms the state-of-the-art Kalman filter based method in terms of both speech quality and intelligibility.