Overview: Laser vibration measurement technology has made great progress in the past decades, and there is still a higher demand in the measurement accuracy and measurement range. Due to the large amount of noise in the measurement process of the laser vibrometer, which leads to a decrease of the vibration measurement precision of the laser vimeter, the filtering of the vibration measurement signal is the key to improving the precision of the laser vimeter. Traditional filters such as the FIR and IIR are time-invariant filters, whose parameters are fixed and invariable. The frequency range of the input signal is required to be known during design, and the filtering performance is inversely proportional to the bandwidth. An adaptive filter is a time-variant filter, it does not need to predict the statistical properties of interference noise, can filter in successive iteration processes of the working state of convergence to adaptively based on the optimal solution under the certain standards, such as minimum mean square error (MMSE) and least-squares criterion, are effective for broadband and narrowband noise suppression. The Least Mean Square (LMS) forward predictor is a kind of the classical adaptive filter, which is based on the MMSE criterion and uses the stochastic gradient descent method to approach the optimal solution under the MMSE criterion by iteration. It has the advantages of simple structure and good robustness. Aiming at the adaptive filtering problem in the laser vibrometer, the Least Mean Square (LMS) forward predictor was used in this paper, and the vibration measurement signal model was established. We simulated and analyzed the parameters of the LMS forward predictor, such as the peak value and frequency of the vibration measurement signal, the order of the filter and the step size coefficient on the filtering performance, and built an experimental system for verification. Simulation and experiments show that the LMS forward predictor can be used as a way to realize adaptive filtering of laser vibrometers, which is suitable for vibration velocity signal filtering in applications such as building vibration detection, mechanical vibration measurement, and material surface micro-damage detection. The filtering effect and convergence speed of the LMS forward predictor are affected by the peak value, order, and step coefficient of the input signal. The filter parameters can be selected and designed according to the requirements of the system for the minimum filtering signal-to-noise ratio and vibration velocity measurement range. This paper provides a theoretical basis for the parameter selection of the LMS forward predictor and provides a technical means for designing an adaptive filter suitable for a laser vibrometer.Aiming at adaptive filtering in the laser vibrograph, we simulated and analyzed the parameters of the Least Mean Square forward predictors, such as the peak value and frequency of the vibration measurement signal, the order of the filter and the step size coefficient on the filtering performance. Then we built an experimental system for verification, and an experimental system was built to verify it. The research results can be used as the theoretical basis for the parameter selection of the minimum mean square error forward predictor and provides a technical means for designing an adaptive filter suitable for laser vibrometer.