Abstract — Noise causes a deterioration of quality against the original information signal. To get a signal that is free from noise disturbance is required an adaptive filter, with the adaptive filter comparison of Peak Signal to Noise Ratio (PSNR) is expected to become larger after the filtering process. The Kalman filter is one of the adaptive filters that can provide efficient calculations for minimizing the noise in the sound signal. In this study, the noise inputs used were wind noise, cheer noise, rain noise, road noise, market noise, and school noise. After simulating Kalman filter is executed then measured signal quality with the main indicator PSNR for each noise input. In this study, the largest PSNR value in noise input cheer with a value of 69.6788 dB, for the smallest PSNR value obtained at 8.4181 dB in market noise input. Noise suppression is best done in market noise input with an increase in PSNR value of 13.7910 dB. Noise suppression performed by Kalman filters works well from the PSNR value after filters and signal graphs and graphs from the spectrograms.