The present study is focused on the detection of certain skin diseases in dogs by using the ResNet-50 neural network architecture. The main objective of the research is to improve the diagnosis of skin diseases in dogs. The problem addressed lies in the need for efficient and reliable detection of dermatological diseases in dogs, which represents a challenge for veterinary professionals, due to the similarity of various skin diseases. The methodology used is based on the convolutional neural network model ResNet50, pre-trained with a large number of images of different categories. The model is specifically adapted to identify and classify canine skin conditions, these are dermatitis, pyoderma and scabies. A transfer learning technique is used for this purpose, taking advantage of the knowledge previously acquired by the model. The PetNet model based on Resnet50 has achieved 85% accuracy in classifying skin diseases in dogs. Similarly, individually PetNet has achieved 81%, 84% and 92% for detecting dermatitis, pyoderma and scabies, respectively. These results support the usefulness of the proposed model as an effective tool for the diagnosis of skin diseases in dogs. In summary, this research presents an innovative approach for the detection of skin diseases in dogs by implementing the PetNet model based on Resnet50. The results obtained reflect a high accuracy in the classification of canine skin conditions, which can significantly contribute to improve veterinary care and improve the quality of life of dogs.