A new system to optimize both control rod pattern and fuel-loading design in boiling water reactors is shown. The system is named OCOTH, and it is based on heuristic techniques such as genetic algorithms, neural networks, and ant colonies. Each heuristic technique is used to design a part of the optimization process. So, the neural network finds an initial fuel loading with a Haling burnup calculation. The ant colony system optimizes full-power control rod pattern of the initial fuel loading. Finally, the genetic algorithm optimizes fuel loading with the optimized control rod patterns. The ant colony system and the genetic algorithm perform an iterative loop until a stop criterion is fulfilled; for example, control rod pattern and fuel-loading convergence. The OCOTH system was tested in an equilibrium cycle of Mexico's Laguna Verde Nuclear Power Plant. We found very good results in control rod pattern and fuel-loading coupled optimization.