Estimation of critical load (CL) is important for soil environmental management and pollution prevention. We developed a mass balance-based dynamic critical load (DCL) model, which improved the model performance, applicability and functionality compared with the traditional one. Paddy soils in two typical fields in central south China and two scenarios were chosen as case studies. The result of case study showed that atmospheric deposition was the main source of Cd, Cu, Pb, and Zn in the soils, with percentage contributions ranging from 59.9 to 79.8%. Crop uptake, particularly the rice straw harvest, was the primary output pathway, accounting for 35.1-71.2% of the total output flux. The critical loads also known as annual input limits (I) of heavy metals in the paddy soils were calculated by the developed DCL model. For example, the Imax of Cd was recommended as 0.05 kg ha in the paddy soils under the default scenario for a protection period of 40 years, and that became 0.12 kg ha and 0.17 kg ha under the straw removal scenario in the two typical fields, respectively. The scenario simulation suggested that the straw removal strategy reduced the total concentrations of heavy metals (C) in the soils and notably increased the I. Meanwhile, the sensitivity analysis indicated that the changes of C and I can be controlled by adjusting the partition coefficient (K), plant uptake factor (PUF) and input flux. The mass balance-based DCL model provides a reference method to establish the standard for controlling heavy metal inputs to agricultural soil, this will be helpful to develop strategies for the prevention of soil contamination.