The commercial release of affordable, low-cost, and consumer-ready virtual reality (VR) devices has increased the accessibility for researchers to investigate the benefits of VR technology including those aimed at education and training. VR technology provides several opportunities that may provide benefits over traditional training methods, this is particularly relevant for safety training due to its ability to safely simulate dangerous scenarios that would otherwise be difficult to access. When implementing a new technology, it is important to evaluate and validate its effectiveness. This paper presents a systematic review of VR safety-relevant training studies that perform an evaluation of their effectiveness. This comprehensive review includes 136 studies published between 2016 and August 2021. Results presented in this paper include application domains, study objectives, study designs, and evaluation measures. Results show that the majority of studies were applicable to health services with the majority focusing on effectiveness evaluation using true- or quasi-experimental design. This study then categorizes each reported evaluation measure into one of the four levels in Kirkpatrick’s model for training evaluation, results showed that the majority of studies evaluated learning (72.06%) and reaction (66.18%) levels with very few studies evaluating behavior and results levels. This study concludes by providing insights and recommendations to help future researchers make informed decisions when designing an effectiveness evaluation study for VR safety-relevant training applications.