Human-induced initiators (category-B actions) are the initiators that are caused by human errors and are rarely explicitly identified and modeled in probabilistic safety assessments (PSAs). The current concern over the safety of multi-unit nuclear power sites is also a motivation for this research. This study proposes a novel process for identifying and quantifying category-B actions and ultimately, how to derive a human-induced initiating event frequency in a multi-unit scenario. Hence, this study fundamentally applies a scenario–system–action search scheme using maintenance and testing procedures, quantifies the human error probability by using the cause-based decision tree and technique for human error rate prediction method, models category-B human actions in the developed fault trees, and derives the human-induced initiating event frequency. The procedure, which is used in this approach, essentially involves system analysis, fault tree development, human error identification, screening, and quantification. The multi-unit loss of offsite power is used as an example accident situation which illustrates the application of the suggested method. Hence, the human-induced initiating event frequency for the loss of off-site power scenario for two units is derived. The application of this method would advance the efforts concerning multi-unit nuclear power plant (NPP) site risk analysis.