Background: In the past 60 years, Cannabis sativa L. has been an object of increasing interest because of the psychotropic effects of some of its constituents. These effects mainly arise from the cannabinoid ∆ 9 -tetrahydrocannabinol (∆ 9 -THC). C. sativa species also synthesize and accumulate the non-psychotropic compound cannabidiol (CBD). Due to their therapeutic potential, both cannabinoids are an object of medical research and drug development. More recently, CBD has received increasing interest as an ingredient in electronic cigarette liquids (e-liquids). This trend may have been reinforced by health and disease-related claims, often based on clinical studies, which are used to advertise CBD. CBD liquids may be based on full-spectrum hemp extracts, CBD isolates, or synthetic CBD, all of which may contain some residual levels of ∆ 9 -THC from either natural content (in the extracts) or from possible degradation of CBD to ∆ 9 -THC, which may occur during storage. There is uncertainty about safety regarding the consumption of CBD (and ∆ 9 -THC) in e-liquids. The aim of this publication was to present an approach for a toxicological risk assessment of CBD and ∆ 9 -THC relevant to e-liquids by using the benchmark dose (BMD) approach. Materials and Methods: Before an analysis to estimate a reference dose (RfD) for both cannabinoids, a systematic review of dose-response data was conducted. The data obtained were analyzed using the BMD approach to derive a benchmark dose lower confidence limit (BMDL). The BMDL was used as a point of departure to estimate the RfD. Results: No adequate human data suitable for dose-response modeling were identified. Based on animal data, the RfD values for the most sensitive endpoints were selected. For CBD, an RfD for acute exposure of 1 mg/kg body weight (bw) was estimated. For ∆ 9 -THC, an acute RfD was found to be 0.006 mg/kg bw. Additionally, the RfD for chronic exposure to CBD was estimated to be 4 mg/kg bw per day. The respective endpoints for CBD were a reduction in norepinephrine turnover and a reduction in uterus weight. The endpoint for ∆ 9 -THC was a change in blood pressure. Conclusions: Because of the limited availability and quality of dose-response data, it cannot be excluded that the estimated RfD values might be afflicted with considerable uncertainties. Therefore, it is recommended to conduct further research on dose-response data, preferably from human studies.