The Teaching Learning-Based Optimization (TLBO) algorithm is a popular meta-heuristic search algorithms inspired by human social interactions. It is popular because of its competitive converging speed and high accuracy, but it has certain limitations. The original TLBO converges to the local optimum in many instances. In this chapter, an Improved Teaching Learning-Based Optimization (ITLBO) method is proposed for solving a classical discrete NP-complete combinatorial optimization 0-1 Knapsack Problem (KP01). Hence, a mathematical model of Improved Penalty Function is developed and used to compute objective function of KP01 efficiently. Accordingly, the proposed algorithm can find a global optimum solution with a better convergence rate. The proposed ITLBO is tested with certain benchmark KP01 instances available in the literature. The results obtained from ITLBO algorithm are compared with some other meta-heuristic algorithms.