Water scarcity has been a critical constraint to economic development in semi-arid areas of China, so optimizing irrigation scheduling has become essential. This study obtained quantitative relationships between crop yield, crop water consumption, and irrigation quantity based on the Hydrus-2D and Stewart models. Different irrigation scheduling scenarios were evaluated to obtain the best irrigation scheduling based on the principle of simultaneous water conservation and crop productivity improvement with the evaluation indicators of crop yield, water use efficiency (WUE), irrigation water use efficiency (IWUE), and Critic-Topsis method. Taking drip irrigation under mulch as an example, the problem of optimizing the irrigation scheduling for different typical years was calculated. The optimization results showed that in the wet, normal, dry, and very dry years the annual irrigation quantity should be 49.68 mm, 49.68 mm, 85.38 mm, and 123.72 mm, when the WUE as well as IWUE, increases significantly, which had less impact on the crop yield and can save irrigation quantity by 30.00%, 30.00%, 35.00%, 27.00%. This study used Hydrus-2D to make a new attempt in irrigation scheduling optimization, giving full play to the model's high accuracy in soil water transport simulation and flexibility in boundary condition simulation. The optimization results can provide a reference for achieving accurate control of irrigation quantity during the crop growth period and reasonable irrigation scheduling formulation for regional crops.