In this groundbreaking study, artificial neural networks (ANNs) are employed to predict the production cross-sections of crucial radioisotopes, namely 18 O, 209 Bi, 232 Th, and 68 Zn, via the (p,n) reaction. We employed a comparative approach to validate the ANN model's predictions by comparing them to outputs generated by established nuclear reaction codes (TALYS 1.9, EMPIRE-3.2 (Malta)) and data from the authoritative source, the Experimental Nuclear Reaction Data (EXFOR).Motivated by the increasing demand for radioisotopes in precise medical diagnostics and successful therapies, this study focuses on investigating methods and new techniques for determining production cross-sections with high accuracy, which are crucial for the consistent supply of vital radioisotopes. In line with this objective, the ANN model demonstrated exceptional performance, achieving remarkably high correlation coefficients, exceeding 0.999 for training and all data, and reaching 0.98665 for testing. Supportive of this, the high correlation coefficients indicate that the ANN estimations effectively match experimental data. Significantly, our findings illustrate the potential of ANNs as a promising alternative for estimating the production cross-sections of 18 O, 209 Bi, 232 Th, and 68 Zn, with the possibility of extending this application to other medically relevant radioisotopes.