Abstract: The lot equipment under the environment of single channel narrow-band signal blind source separation, it is of great significance in enhancing channel balance. Traditional blind source separation algorithm USES the quadrature amplitude modulation channel allocation method, on the Internet of things more equipment environment about the worse performance of narrowband signal blind source separation problem. Based on double threshold receives the single channel of narrowband signal blind source separation algorithm is proposed. Build the double threshold decision variable signal analysis model, design of single channel narrow band correlation processor, based on fast Fourier transform calculation method of the single channel narrow band correlation processing, using piecewise fitting method to compensate the nonlinear distortion, to control the phase information of the iterative equalizer, in combination with double threshold judgment receiving algorithm, implements the narrowband signal blind source separation algorithm is improved. Studies show that the proposed algorithm can effectively improve the narrowband signal blind source separation detection performance, the effective implementation of the original signal recovery and feature detection, improve the channel gain and achieve channel equalization.