We provide an algorithm with a higher order convergence than New-ton’s method. The algorithm converges at order k and has a controlled error bound. We expect that the advantages of the new algorithm tobe more significant when extreme accuracy is required.