Diabetes alert dogs inform diabetic owners when the owner has experienced a potentially dangerous drop in blood glucose levels. It is believed that the dogs alert to a change in the odour of the owner’s breath during these episodes. However, training these dogs is expensive, thereby making this resource unavailable for many diabetic patients who could benefit from the presence of a diabetes alert dog. The aim of our study was to determine whether pet dogs can be trained to reliably alert to the presence of a scent (myrrh) on their owner’s body within a 10 week training period. We recruited 18 pet dogs and their owners in the Greater Melbourne, Victoria, Australia, region to participate in a 10 week training program with a hobby scent trainer. After the 10 week training program, dogs and owners participated in a 4 week testing session, in which the owners wore either a waterproof container containing myrrh (scent-present), or an identical waterproof container that did not contain myrrh (scent-absent). Occasionally, the owner would open the waterproof container to release the scent. We measured when the dog alerted to the presence of the scent when the scent-present container was open (true positive) or closed (false positive), and when the dog did not alert when the scent-present container was open (false negative). We also compared these results to the same information when the owner was wearing the scent-absent container, in order to control for visual and auditory cues from opening and closing the container. If dogs provide more true positives than false positives or false negatives, it will provide evidence in support of a 10-week training requirement for pet dogs. This may ultimately permit diabetic owners to train their own dog to alert to a change in their blood glucose levels.