A new search method for box-constrained optimization problems titled Search Method for Box Optimization (SMBO) is presented in this paper. SMBO is a population heuristic based method intended to solve box constrained global optimization problems. SMBO represents the population as Probability Density Functions (PDF) within the problem bounds. The PDF shape is dynamically adapted during the search process leading to convergence towards the global optimum. The method is tested on two benchmark sets, which include unimodal, multi-modal, expanded and hybrid composition functions. The performance of SMBO is compared with several genetic algorithms (GAs); the first test compares it with relatively traditional/classic nine codes of parallel GAs, and the second compares SMBO with two recent variants of GAs. The obtained results show equal or better performance in both comparisons. The method has also been applied to optimize a nonlinear model for management of a water supply system, and the results were compared with the commercial GA toolbox of MATLAB.