Alkaline electrolysis (AEL) is the most mature electrolysis technology and is widely used in large-scale power-to-gas scenarios. However, the loading range of AEL system is restricted by the minimum load, 10%–40% in general, which prevents system operation during low-load periods. This minimum load is caused by the flammable mixture formed by the impurity crossover and is related to operation conditions such as current and pressure. In this paper, we propose a novel pressure control strategy to extend the loading range of an AEL system by reducing the system pressure at low-load periods while also maximizing the system efficiency. A dynamic impurity accumulation model is established illustrating the flexibility limitation mechanism of the AEL system. Two pressure controllers are designed based on the steady-state and dynamic impurity accumulation model, including an operation curve tracking controller (OCT) and a model predictive controller (MPC). In the case study, both controllers achieve satisfying results under the peak shaving scenario, with an extension of the AEL system's minimum load from 27.5% by the traditional constant pressure controller to 10.5% by the OCT controller and 10% by the MPC controller. This pressure control strategy reduces the electricity abandonment and improves the system economy significantly.