Context: Fire and controlled grazing have been widely adopted as management interventions to counteract woody shrub proliferation in many arid and semiarid grassland systems. The actual intensity of grazing and fire, along with the timing of the interventions, however, are difficult to determine in practice. Objectives: This study aims to establish model simulations to access the long-term landscape changes under different land management scenarios. Methods: We developed a cellular automata model to evaluate landscape dynamics in response to scenarios of grazing, fire, time of intervention, and initial coverage of grasses and shrubs. Results: With current grazing intensity and fire suppression, the landscape may shift to a shrub-dominated landscape in 100–150 years. An appropriate combination of grazing and fire management could help maintain over 50% of grass cover and reduce the shrub cover to less than 2%, keeping the landscape highly reversible. Even using 1% grazing intensity and periodic fire once a year, the management tools should be implemented in 60 years, otherwise, they may lose effectiveness and the vegetation transition to grasslands would become impossible. Conclusions: This study highlighted that the reintroduction of fire not only directly removes shrubs but also reallocates soil water and resources among different microsites, which may accelerate grass recovery and suppress shrub regrowth, potentially reversing the shrub invasion process. The combined grazing and fire management plans should be carried out before a threshold time depending on the chosen management tools.