Particle/air jet flow has been widely used in various fields such as drying powder treatment and internal combustion engines. This study was conducted to model the relationship between the turbulent velocity field at the upstream portion of the particle/air jet and the particle concentration distribution at the downstream portion using an artificial neural network. The turbulent velocity field and the particle concentration distribution were simultaneously measured by applying the particle image velocimetry (PIV) system with three lasers to both the upstream and downstream portions of the particle/air jet flow. An approximate model for predicting the particle concentration distribution at the downstream portion from the turbulent velocity field at the upstream portion was constructed by the neural network using measured data. The neural network actually yielded a precise approximation by applying the moving average method to the turbulent velocity data and by suppressing the noise.