Aiming at the noise caused by human interference and equipment defects in the process of laser scanning, a steady-state filtering technology of spectral data based on the informatization of laser scanning terminal is proposed to effectively remove the noise. According to the composition structure of the laser scanning system and the information process of the terminal, the noise source in the scanning process is determined, and the spectral data noise model is established under the influence of factors such as beam divergence, scanning angle and equipment defects. After wavelet transform, the initial signal is decomposed into low frequency and high frequency. Soft threshold function and hard threshold function are integrated to describe the changing trend of spectral data after wavelet decomposition, and the filtering threshold function is set to realize noise filtering. Simulation results show that the proposed technique can effectively improve the signal-to-noise ratio of spectral signal, and the root mean square error between the filtered signal and the real signal is the smallest.