A Least Mean Squares (LMS)-style algorithm is derived for the feedback control problem. The algorithm allows a tap delay line within the closed loop to be used for control applications. This paper derives the algorithm and applies the algorithm to two simple control problems for which the solution is known and to one problem with an unknown solution. The first problem is a stable second-order system. The second problem is a unstable second-order system which is initially stabilized with the feedback loop. In both problems, the weights converge to the expected values. The stable problem is used again with an inaccurate model that has 50% more damping than the actual plant. The weights converge to a solution which increases the performance of the controller.