The invention discloses a processing method of interference signals in decoding process of a digital coder, which can effectively filter an interference signal with smaller width in pulse signals A and B by inducing one external clock signal the frequency of which is more than 3 times, but less than 10 times of the maximum frequency of the pulse signal A or B generated and outputted by the digital coder and taking the external clock signal as the clock signal of each-grade latch of a first digital filter and a second digital filter, and can guarantee not to change phase position factors of the pulse signals A and B; while for the interference signal with larger width, if the pulse signal B is high level when the pulse A is jumped from high level to lower level, then the attribute value isincreased by 1; however, if the pulse signal B is still high level when the interference signal stops and the pulse signal A is jumped from low level to high level, then the attribute value is reduced by 1, thus being capable of effectively eliminating the influences of the interference signal with larger width on the pulse signals, thereby leading a final attribute value to be more accurate.