About one third of all orchid species are deceptive, i.e., not providing any reward to their pollinator. Such species often have lower visitation rates compared to rewarding relatives. This could result in lower levels of geitonogamous selfing and thus would provide an advantage in term of progeny fitness through inbreeding avoidance. This hypothesis could be tested by comparing the level of inbreeding depression between deceptive and rewarding orchids. However, due to the difficulty to raise orchids from seeds, few studies of inbreeding depression are available, and most are focused on very early life stages, such as seed mass or embryo viability. Here, we present the results from an experimental investigation of inbreeding depression in the deceptive flower-colour dimorphic Dactylorhiza sambucina, from in vitro cultivation to greenhouse soil transplantation. We found strong inbreeding depression at all recorded stages (i.e., germination and survival), with estimates ranging from 0.47 to 0.75. Our study finally proposes a simple and suitable experimental protocol to raise orchids from seeds with high germination rates.