This paper presents the Digital Hormone Model (DHM) as a new computational model for self-organization. The model is a generalization of an earlier distributed control system for self-reconfigurable robots [1-3] and an integration of reaction-diffusion model with stochastic cellular automata. In this model, cells secrete hormones, and hormones diffuse into space and influence the behaviors of other cells to self-organize into global patterns. In contrast to the reaction-diffusion models, cells movements are computed not by the Turing's differential equations [4] hut by stochastic rules that are based on the concentration of hormones in the neighboring space. Different from simulated annealing, the stochastic rules are not the Metropolis rule (5], which will not produce the desired results here. Experimental results have shown that the simulated cells in the DHM produce results that match to the observations made in the biological experiments where homogenous skin cells selforganize into feather buds regulated by hormones [6]. With the unique properties of being distributed, scalable, robust, and adaptive, the DHM opens opportunities for new hypotheses, theories, and experiments for further investigating the nature of self-organization