This thesis describes the construction and implementation of a microprocessor controlled recursive adaptive filter applied as a noise canceller. It describes the concept of the adaptive noise canceller, a method of estimating the received signal corrupted with additive interference (noise). This canceller has two inputs, the primary input containing the corrupted signal and the reference input consisting of the additive noise correlated in some unknown way to the primary noise. The reference input is filtered and subtracted from the primary input without degrading the desired components of the signal. This filtering process is adaptive and based on Widrow-Hoff Least-Mean-Square algorithm. Adaptive filters are programmable and have the capability to adjust their own parameters in situations where minimum piori knowledge is available about the inputs. For recursive filters, these parameters include feed-forward (non-recursive) as well as feedback (recursive) coefficients. A new design and implementation of the adaptive filter is suggested which uses a high speed 68000 microprocessor to accomplish the coefficients updating operation. Many practical problems arising in the hardware implementation are investigated. Simulation results illustrate the ability of the adaptive noise canceller to have an acceptable performance when the coefficients updating operation is carried out once every N sampling periods. Both simulation and hardware experimental results are in agreement.