Eliciting Emotions through Facial Analysis delves into the interpretation of facial expressions in dogs. Understanding these expressions is crucial for strengthening the dog-human bond and ensuring canine well-being. This paper introduces FaceEvoke, a novel custom CNN model meticulously crafted for the precise detection of changes in eye shape, mouth position, and ear movement. The model seamlessly integrates multi-scale features and convolution, empowering observers to decipher a dog’s emotional state with enhanced precision. The standout quality of this innovative model is its proficiency in accurately categorizing a wide array of dog facial expressions drawn from a diverse image set. Experimental results underscore the superior performance of our model, surpassing other established algorithms with an impressive accuracy rate of 99.87%, thereby affirming its efficacy in veterinary care and animal behavior studies.