In this paper, efficient least mean squares , normalized least mean squares and sign least mean squares algorithm are proposed for removing artifacts preserving the low frequency components and tiny features of the ECG. The proposed implementations are suitable for applications requiring large signal to noise ratios with fast convergence rate. The sign least mean squares algorithm, being the solution of the steepest descent strategy for minimizing the mean squared error in a complete signal occurrence, is shown to be steady-state unbiased and with a lower variance than the LMS and NLMS algorithm. Finally, we have applied this algorithm on ECG signals and compared its performance with the LMS and NLMS algorithms. The results show that the performance of the sign least mean squares algorithm is superior to the LMS and NLMS algorithms.