This work presents a new algorithm to solve the distribution network reconfiguration problem based on the Tabu Search algorithm. The proposed method provides a way to obtain new solutions and neighbors for the Tabu Search algorithm with controlled randomness, different from what is available in the literature. The proposed method manages to make the algorithm produce good solutions in most iterations, thus performing a smooth and constant descent to obtain the final solution. The algorithm seeks to reduce losses and improve the voltage profile in medium voltage distribution systems. Tests were carried out in two systems used in the literature, the 33-Bus and the 94-Bus systems. The simulation results show that the proposed method has a low computational cost and obtains up to 100% optimal solutions in 100 executions.