Spray combustion remains the dominant mode of energy conversion, providing the majority of the world＇s energy requirements. A good understanding of spray formation processes and spray droplet size distributions is essential for the design and operation of spray combustion systems with high energy efficiency and low pollutant emissions. The early stage of the spray formation process is clearly deterministic, with distinct unstable wave motion, whereas the final stage of spray formation process is more or less random, chaotic, and stochastic due to nonlinear effects of the unstable wave development. The number of droplets produced in a spray is enormous, and the description of each individual droplet becomes highly improbable, thus requiring a statistical treatment. The present model incorporates the deterministic aspect through the linear and nonlinear stability theory, and the stochastic aspect through the maximum entropy principle. It can predict, from a given flow condition at the nozzle exit, the spray formation process and the probability distribution of subsequently formed droplets in sprays. The effect of flow conditions at the nozzle exit on the droplet size distributions has been investigated. The present predictive model gives the initial distribution of droplet diameters and velocities in sprays, and hence will be useful as a submodel for overall spray combustion modeling.