Simulating the behavior of ant colony searching for food, the Ant Colony Algorithm is an emulated and optimized algorithm, which demonstrates excellent performance in such combinatorial optimization problems as Traveling Salesman. However, it has not been widely applied to the inversion and interpretation of magnetic data. Based on the characteristics of magnetic data inversion and interpretation, this paper improves the mapping mechanism from objective function value to pheromone and summarizes Ant Colony Algorithm optimizing for continuous and multiple objective function using nodes partition strategy. Satisfactory results were achieved when Ant Colony Algorithm was simulated to inverse the parameters in the synthetic model experiment of magnetic data and was applied to prospect the banded iron formation according to low-altitude aeromagnetic data surveyed at Iron Mount mining area, southern Australia.