In the last two decades, recreational drug use patterns have changed substantially, particularly with opioids and cannabinoids. Opioid-related deaths continued to rise, with interventions often making the problem worse. Attitudes towards cannabinoids were quickly changing, causing a trend of increased cannabis use and legalization. The impact of these drugs on human populations were and remain an important public health concern. Although these drugs may also be harmful to vulnerable populations, such as pet dogs, their impact on dog populations had received little attention. As the basic epidemiology of opioids and cannabinoids poisonings of dogs has not been established, we used data concerning reports of dog poisoning events from the American Society for the Prevention of Cruelty to Animals’ (ASPCA) Animal Poisoning Control Center (APCC), from 2005–2014, to begin building this foundational knowledge in US dogs. We combined the APCC data with several other data sources and used multilevel multivariable logistic regression models to identify legislative, drug use patterns, socioeconomic, and dog-level variables associated with opioid and cannabinoid poisonings in dogs. We identified clusters of calls concerning cannabinoid and opioid poisonings in dogs in space, time, and space-time using scan statistics. We fit ordinary and mixed logistic regression models to identify pet iii demographic and health disorders associated with opioid and cannabinoid dog poisonings, and compared their predictive performance to analogous lasso logistic regression models. This study revealed the need for increased awareness among the pet owning public, and public health and veterinary communities about the impact of dog-level and regional factors on the risk of dogs being poisoned with these toxicants, and the value of improved surveillance of these events.