This study presents an adaptive parameter control structure for the New Local Branching Algorithm (NLBA) to enhance the overall efficiency of the algorithm by implementing an exploitation and exploration approach within solution search subspaces. In other words, the optimization problem is divided into several sub-problems and the proposed adaptive parameter control structure introduces some deterministic rules to check the convergence speed and adapt the performance of the algorithm in each subspace. So, the proposed adaptive structure uses an exploration approach to avoid becoming trapped at local optimum that may not be the global optimum, as well as an exploitation approach to look near good solutions for even better ones at each iteration. To evaluate the performance of the proposed adaptive parameter control structure, 26 multi-commodity network design problems are tested. The experimental results of this adaptive structure for NLBA (ANLBA) confirm that its performance compared to the original NLBA, a hybrid structure that consists of the Self-Adaptive Harmony Search (SAHS) algorithm as a parameter tuning method for NLBA, and CPLEX is significantly improved.