Background and Objective The risk of thrombotic complications in critical patients with COVID-19 remains extremely high, and multicenter trials failed to prove a survival benefit of escalated doses of low-molecular-weight heparins (nadroparin calcium) in this group. The aim of this study was to develop a pharmacokinetic model of nadroparin according to different stages of COVID-19 severity. Methods Blood samples were obtained from 43 patients with COVID-19 who received nadroparin and were treated with conventional oxygen therapy, mechanical ventilation, and extracorporeal membrane oxygenation. We recorded clinical, biochemical, and hemodynamic variables during 72 h of treatment. The analyzed data comprised 782 serum nadroparin concentrations and 219 anti-Xa levels. We conducted population nonlinear mixed-effects modeling (NONMEM) and performed Monte Carlo simulations of the probability of target attainment for reaching 0.2–0.5 IU/mL anti-Xa levels in study groups. Results We successfully developed a one-compartment model to describe the population pharmacokinetics of nadroparin in different stages of COVID-19. The absorption rate constant of nadroparin was 3.8 and 3.2 times lower, concentration clearance was 2.22 and 2.93 times higher, and anti-Xa clearance was 0.87 and 1.1 times higher in mechanically ventilated patients and the extracorporeal membrane oxygenation group compared with patients treated with conventional oxygen, respectively. The newly developed model indicated that 5.900 IU of nadroparin given subcutaneously twice daily in the mechanically ventilated patients led to a similar probability of target attainment of 90% as 5.900 IU of subcutaneous nadroparin given once daily in the group supplemented with conventional oxygen. Conclusions Different nadroparin dosing is required for patients undergoing mechanical ventilation and extracorporeal membrane oxygenation to achieve the same targets as those for non-critically ill patients.