The Index Tracking Problem involves the creation of an investment portfolio that accurately replicates the behavior of a market index. Being an NP-Hard optimization problem, it is well-suited for metaheuristic approaches. Drawing inspiration from the Harmony Search algorithm, the Harmony Search algorithm for Portfolio Optimization is presented. This algorithm addresses the problem’s complex constraints by utilizing two search operators that effectively handle both the problem itself and the most difficult constraints, and it incorporates two population initialization strategies that aid in achieving convergence and can deal with problems of big size.