Seismic data is often contaminated with random noise and therefore may be of poor quality due to the low signal-to-noise ratio. Stochastic noise may not only affect speed reversal and seismic transmission, but also create potential hazards in processing and interpreting its overload due to its effects. There are various methods to reduce the effect of noise on seismic data, and we have other methods that are a combination of these methods that overlap the weaknesses of individual methods. One of the most desirable methods is the machine learning method to weaken Is seismic noise, which in this article we compare several combined methods with machine learning method.