Objective To explore the attributes of medication administration errors, to mine the significant association rules between the medication administration errors and the factors influencing the occurrence of the errors , and to provide references for prevention of such errors. Methods The relevant information of 136 cases of medication administration errors occurred in 3 tertiary hospitals in Shanghai from January 2016 to October 2020 were collected retrospectively.The association rules between all characteristics were mined and analyzed using the Apriori algorithm. Results The characteristics of medication administration errors were described from 4 aspects of nurse trait, time distribution of events, administration route, type of errors and error elements.Among the 136 cases of medication administration errors, 125 were made by nurses.A total of 706 strong association rules were found, and 12 association rules with clinical significance between errors and factors were finally extracted.The rule 1 to 3 revealed the attribute of not using PDA while administering drugs, rule 4 to 8 disclosed the attribute of not following standard medication administration process, rule 9 to 12 exposed negligence associated attribute. Conclusion The association rules between such factors as nurses′ length of service, professional titles, education levels, time distribution of events, departments, administration route and the 3 attributes of events(not using PDA while administering drugs, not following standard medication administration process, and negligence) could guide clinical nurses to identify the risk factors of medication administration errors.Clinical managers and nurses are recommended to seek such measures, as strengthening the concept of correct drug use, ramping up training for nurses at all levels, pushing forward timely human resource allocation, improving information construction, etc., in an effort to prevent the occurrence of medication administration errors.