Nutrient status provides crucial information about the stability and potential recycling capacity of a specifi c ecosystem. To comprehend the nutrient element transport in a typical cotton (Gossypium hirsutum L.) cultivation ecosystem, this study established dynamic compartment models to estimate the contents of four elements (Ca, Na, K, and B) in fi ve compartments during fi ve growth stages. Th e stage-variable parameters were determined using three methods. Pot experiments were conducted for model calibration and validation. More than 90% of Ca, Na, and B accumulated in the soil compartment, whereas K decreased throughout the growth season. Aft er the element proportions were introduced into parameter determination, the models yielded satisfactory estimations because 74 to 79% of the relative errors were <20%. For soil and leaf compartments, satisfactory estimations were obtained (modifi ed Nash-Sutcliff e coeffi cient of effi ciency C eff _m > 0.9), whereas the average C eff _m of root, stem, and fruit compartments were 0.82 ± 0.051, 0.78 ± 0.073, and 0.62 ± 0.051, respectively. Th e pot experiments indicated that cotton yields were much higher (P < 0.05) for the brackish water treatment than for the freshwater treatment, despite vegetative growth being inhibited. With an eff ective method for determining parameters, the compartment models successfully estimated the nutrient-element content in the typical cotton cultivation ecosystem, thereby providing a useful tool for proper management of irrigation, fertilization, and alleviation of salt damage.Abbreviations: CM, compartment model; DW, dry weight; M1 to M3, Methods 1 to 3; T1 to T7, Treatments 1 to 7; C eff _m , modifi ed Nash-Sutcliff e coeffi cient of effi ciency.