In this work, optimization of an air cooling system was carried out by using Lagrange multipliers method, differential evolution and particle swarm algorithms for various temperatures and mass flow rates which are inspiring the total cost of the system. It was aimed to show how differential evolution and particle swarm optimization approximates to the exact fitness values for various iterations. A comparative optimization study considering LM, DE and PSO are executed, results are plotted and tabulated in the results section. Total cost of the system is dependent strongly on varying temperatures and mass flow rates of the system. Best cost values are obtained when air mass flow rates, input and output air temperature difference and cooling tower exit temperatures are lower and water mass flow rate is higher. Differential evolution and particle swarm algorithms have given exactly same results with Lagrange methods for most parameter inputs with easier and less time consuming programming. When the initialization of parameters is not setup properly, both DE and PSO may not give desired values. Thus, it is of importance to setup initial parameters and control the search space. It can be controlled by reinitializing these algorithms when off springs in DE and particles in PSO are out of constraint bounds. DE and PSO can improve the calculation performance of air cooling systems and can be conveniently used for cost optimization of air cooling systems.