The canine influenza virus (CIV) outbreaks have raised concerns as they pose a threat to the health of dogs. The successful construction of a canine influenza (CI) infection model is essential to study the CIV. Here we investigated the pathogenicity of different infectious doses of H3N2 CIV in Beagle dogs. Thirty-seven healthy Beagle dogs were used in the experiment and were infected with 103, 104, 105, and 106 50% egg-infectious doses (EID50). Compared to the dogs in the other three groups, those in the 106 EID50 group presented with obvious clinical symptoms, high virus titer, and typical pathological changes. Considering the ensemble of clinical scores, body temperature, virus shedding, lung lesions, pathological section scores, and visceral virus titers, we determined that 106 EID50 is the minimum infectious dose for the Beagle infection model. The other three infectious doses had almost no clinical symptoms. These results indicate that 106 EID50 is the minimum infectious dose of H3N2 CIV that can cause obvious clinical manifestations and pathological changes associated with CI in Beagle dogs. The theoretical framework developed in this research will guide the establishment of an infection model of CIV for future research.