The classical ecological theory of patch dynamics and community structure emerged in the early 1900s as a way to reconcile persistent patterns observed in the overall structure and function of forest and plant communities with continual change at the local or patch scale. The original theory was purely conceptual, but mathematical formulations based on Markov-chain models with empirical one-step transition probabilities began to appear in 1970. The main purpose of the present study is to incorporate three patch-scale ecological processes (succession, retrogression, and disturbance) into a Markov-chain model of patch dynamics, yielding a stochastic model that links patterns at the community scale to ecological processes acting at the patch scale. The main results deal with effects of the three patch-scale processes on the form of the stationary patch-state distribution and the length of time required to achieve approximate stationarity. The “forward” process of succession tends to concentrate patches near the ultimate successional state, while the “backward” processes of retrogression and disturbance tend to concentrate patches near the initial successional state; the form of the stationary patch-state distribution reflects the balance or imbalance between these processes. Frequent disturbance also tends to shorten the time required to achieve approximate stationarity.