A method for optimally simulating fluid flow around a real object by (a) defining an initial state of a simulation space having a plurality of lattices with nodes; determining a space occupation state by objects, a fluid state, and a probability distribution state of a particle for respective nodes; generating a flow effect for the respective nodes using a collision rule which is a probability distribution of the particle with respect to the velocity thereof, obtained by satisfying a condition at which a Tsallis or a Rényi-divergence between the collision rule itself and a reference collision rule provides a simulated fluid flow result; and renewing the initial state and varying the space occupation state based on the simulated fluid flow result; and (b) based on step (a) modeling fluid flow around the real object.