The process of denoising input speech signals is more helpful in the process of providing efficient sound system. Fast adaptive Kalman filter is designed for the removal of the noises in the signals. Kalman filter includes two steps the prediction step and the estimation step. In the prediction step the noises in the signal were estimated. In the estimation step the estimated noises in the images were removed based on the calculation of the amount of noise estimated. Fast adaptive Kalman filtering is employed for the removal of the noises from the signal which is based on the prediction and estimation of the noise level in the signal. The input speech signal is denoised with the help of the Fast adaptive Kalman filter. State transition and observation models need not be linear functions of the state but may instead be non-linear functions. The function can be used to compute the predicted state from the previous estimate and similarly the function can be used to compute the predicted measurement from the predicted state. The prediction and the estimation process are the steps in the kalman filter. The nonlinear functions employed for the prediction and estimation process improves the performance of the adaptive kalman filter process. The non-linear functions defined for the kalman filtering process is designed so that the process is reduced in iterations and more adaptive. The signal subspace algorithm which identifies the maximum level of the noise in the input signal and the removal of the identified noise level is employed. The filtered signal and the original signal is then compared inorder to measure the performance of the process. The performance is measured with the help of performance metrics like SNR, DRR, LSD, SRMR and WSIR-IN.