OBJECTIVE： To mine the signals of osimertinib adverse drug event （ADEs） and to provide reference for clinical drug safety. METHODS： By reporting odd ratio （ROR） and Bayesian confidence propagation neural network （BCPNN） algorithms， data mining and signal detection were carried out on osimertinib-related ADEs reports with case number ≥3 reported by American Adverse Event Reporting System （FAERS） from Nov. 2015 to Jun. 2018. RESULTS & CONCLUSIONS： A total of 2 044 osimertinib-related ADEs were found， among which the female （1 285 cases） was more than the male （615 cases）， mainly aged 45-64 years （228 cases）， 65-74 years （241 cases）， ≥75 years （248 cases）； most of reporters were physicians （887 cases）. A total of 63 and 57 osimertinib ADEs signals were mined by ROR and BCPNN methods， and the latter signals overlapped with the former. The obtained signals included skin toxicity （dermatitis acneiform， rash， dry skin， onychoclasis， nail disorder， paronychia）， gastrointestinal tract reaction （decreased appetite， diarrhea， stomatitis）， respiratory diseases （interstitial pneumonia， pneumonia）， hematological diseases （platelet count decreased， white blood cell count decreased， neutrophil count decreased）， vascular events （pulmonary embolism， deep vein thrombosis）， electrocardiogram QT prolonged （RORs were 2.04-56.70， ICs were 0.97-4.43）， all of which had been recorded in drug instructions； obtained signals also included new suspicious signals not indicated in the drug instructions， such as liver damage， cerebral infarction， hyponatremia， atrial fibrillation， renal impairment， dehydration， dysgeusia， and so on （RORs were 2.06-161.74， ICs were 1.00-4.58）. In clinical use of osimertinib， besides paying close attention to known safety problems such as nail toxicity of finger （toe） and rash， interstitial pneumonia， electrocardiogram QT prolonged， we should also be alert to the occurrence of potential ADEs such as liver damage to ensure drug safety.