As a positron-emitting nuclide that can be co-labeled with therapeutic nuclides such as 47Sc and 177Lu, 44Sc holds significant promise for applications in theranostics in clinical nuclear medicine. To optimize its production, it is essential to simulate the nuclear reaction processes using Monte Carlo methods in order to obtain optimal experimental parameters. In this study, we employed the TALYS-2.0 program package to calculate the reaction cross sections of several nuclear reactions that may occur in the actual process. The stopping power of CaCO3 targets for protons was evaluated using the SRIM program. Additionally, FLUKA was utilized to simulate nuclide production in a cyclotron, allowing us to investigate the influence of various experimental parameters, including target thickness, incident proton energy, irradiation time, and cooling time, on the yields of 44Sc and co-produced radionuclide impurities. The results show that choosing thinner or thicker rather than medium thickness is more favorable for obtaining higher radionuclide purity of 44Sc. Incident proton energies in the range of 13-14 MeV can produce higher activity of 44Sc. Longer irradiation time, preferably 8-12 h, results in higher activity and higher radionuclide purity of 44Sc. A relatively longer cooling time, e.g., 1-2 h, can effectively reduce the activity of radioactive impurity nuclides, such as 13N and 18F, and significantly increase the radionuclide purity of 44Sc in the final product. This study also verified the accuracy and reliability of the calculations by comparing the results of different programs, as well as comparing the results of the calculations in this study with the experimental results of previous researchers.