•This paper provides a systematic review of researches on FMEA improvement.•A set of 236 journal articles published between 1998 and 2018 was identified.•Bibliometric analysis was conducted to provide insights into the research theme.•Research gaps and opportunities on the improvement of FMEA are identified. Failure mode and effects analysis (FMEA) is a reliability management technique commonly utilized in various industries to guarantee the security and reliability of systems, services and projects. Nonetheless, the classical risk priority number (RPN) method has been criticized for many inherent deficiencies in the literature. Over the last decades, plenty of models and approaches have been employed to enhance the inherent characteristics of FMEA, but few contributions are devoted to review and summarize the related researches on FMEA improvement. Therefore, this paper aims to conduct a systematic review of the journal articles on the topic during the years between 1998 and 2018. A metadata statistical analysis is undertaken to present an overview of publication distribution across time and journals. Besides, a bibliometric analysis is performed to identify the most influential authors, institutions and areas, reveal the research hotspots and give an insight into the theme evolution in this field. The results indicated that the annual publications on FMEA improvement are rising quickly in the past two decades, especially after 2013; Liu HC, Chang KH, Kumar D, Sharma RK and You JX are the most prolific researchers; and Asian especially China is the major contributor in this field. Moreover, “healthcare failure mode”, “risk ranking”, “extended FMEA”, “gray theory”, “risk evaluation”, and “fuzzy inference” are the research focuses in improving the traditional FMEA.