The increasing use of glucagon-like peptide-1 receptor agonists (GLP-1 RAs) highlights the importance of adverse event monitoring as part of pharmacovigilance. Conventional analyses generally focus on event frequency, leaving transition patterns among adverse events within individual reports insufficiently explored. This study aims to analyze adverse event reporting transition patterns associated with GLP-1 RAs using process mining based on OpenFDA Harmonized Fields. Data were obtained from the OpenFDA Drug Adverse Event API, sourced from the FDA Adverse Event Reporting System (FAERS), covering the 2023–2025 period for semaglutide, liraglutide, dulaglutide, and tirzepatide. A total of 188,734 reports were extracted and processed through cleaning, normalization, deduplication, and event log construction, resulting in 496,212 event records from 188,247 cases. The 25 most frequently reported adverse events were selected for analysis, with 48,429 cases containing at least two events. Directly-Follows Analysis identified Nausea → Vomiting as the transition with the highest frequency, occurring 4,459 times. For tirzepatide, the transition Incorrect dose administered → Injection site pain occurred 2,267 times. Comparison of the Heuristics Miner and Inductive Miner showed that the Heuristics Miner achieved precision values of 0.9671–0.9909 and F-scores of 0.8474–0.8682, whereas the Inductive Miner achieved a fitness of 1.0000 with precision values of 0.5315–0.7481. Based on the balance of evaluation metrics, the Heuristics Miner was used as the primary model for transition analysis. The results demonstrate that process mining can be used to explore the structure of adverse event reporting transitions associated with GLP-1 RA therapies.