Data Mining is the most dealt and emerging research area for extraction of useful data nowadays. This is due to the growth of increasing demand of digitization of data in various fields such as bank, wide use of Internet, etc. Data Mining includes main tasks as clustering, classification, rule mining and regression. Further optimization techniques are required so that accuracy is better as compared to existing methods. Optimization is mainly achieved by Swarm Intelligence (Computational Intelligence) which is a part of Soft Computing branch. Various swarm based techniques like Ant Colony Optimization (ACO), Imperialist Competitive Algorithm (ICA), Harmony Search (HS), etc are compared using 4 benchmark function as Sphere, Ackley, Rastrigin and Rosenbrock. The results conclude that Firefly algorithm(FA) has better optimization compared to others. Twelve optimization techniques are compared on basis of these 4 benchmark function.