Myelin Oligodendrocyte Glycoprotein Antibody-Associated Disease (MOGAD) is commonly misdiagnosed as Multiple Sclerosis (MS) in many patients; therefore, this study aimed to differentiate the two diseases using an MRI algorithm. 406 MRI scans were obtained from 19 different centers; these scans included patients with non-acute MS and MOGAD. These scans were analyzed by 2 different readers to test the validity of the MRI algorithm and probability attention maps (PAMs) were drawn to understand the key differences between MS and MOGAD in patients. The validation set illustrated the capacity of the MRI algorithm as it differentiated MOGAD and MS with 75% accuracy. The accuracy increased to 86% when the MRI algorithm was combined with a DL model. PAMs also highlighted the corpus callosum, precentral gyrus, right thalamus, right cingulate cortex, and other structures as key structures for identifying MOGAD. With this technology, physicians can diagnose MS correctly from MOGAD and provide better treatment options for patients.