Woody weeds pose significant threats to the 12.3 billion dollar Australian grazing industry. These weeds reduce stocking rate, increase mustering effort, and impede cattle access to waterways. Two major concerns of woody-weed management are the high cost of weed management with respect to grazing gross margins, and episodic seedling recruitments due to climatic conditions. This case study uses a Stochastic Dynamic Programming (SDP) model to determine the optimal weed management decisions for chinee apple (Ziziphus mauritiana) in northern Australian rangelands to maximise grazing profits. Weed management techniques investigated include: no-control, burning, poisoning, and mechanical removal (blade ploughing). The model provides clear weed management thresholds and decision rules, with respect to weed-free gross margins and weed management costs.