Drug-resistant epilepsy is a common complaint in dogs and affects up to 30% of dogs with idiopathic epilepsy. Experimental data suggest that targeting cyclooxygenase-2 (COX-2) mediated signaling might limit excessive excitability and prevent ictogenesis. Moreover, the role of COX-2 signaling in the seizure-associated induction of P-glycoprotein has been described. Thus, targeting this pathway may improve seizure control based on disease-modifying effects as well as enhancement of brain access and efficacy of the co-administered antiseizure medication. The present open-label non-controlled pilot study investigated the efficacy and tolerability of a COX-2 inhibitor (firocoxib) add-on therapy in a translational natural occurring chronic epilepsy animal model (client-owned dogs with phenobarbital-resistant idiopathic epilepsy). The study cohort was characterized by frequent tonic–clonic seizures and cluster seizures despite adequate phenobarbital treatment. Enrolled dogs (n = 17) received a firocoxib add-on therapy for 6 months. Tonic–clonic seizure and cluster seizure frequencies were analyzed at baseline (6 months) months during the study (6 months). The responders were defined by a substantial reduction of tonic–clonic seizure and cluster seizure frequency (≥50%). In total, eleven dogs completed the study and were considered for the statistical analysis. Two dogs (18%, 2/11) were classified as responders based on their change in seizure frequency. Interestingly, those two dogs had the highest baseline seizure frequency. The overall tolerability was good. However, given the low percentage of responders, the present data do not support an overall considerable efficacy of COX-2 inhibitor add-on therapy to overcome naturally occurring phenobarbital-resistant epilepsy in dogs. Further translational evaluation should only be considered in the canine patients with a very high baseline seizure density.