Nowadays, Skin diseases are become the most infectious diseases occurring in cattle of all ages and often found in humans and animals. Cattle skin diseases are particular kinds of diseases caused by tumors, allergies, parasites, autoimmune, bacteria infectious, and viral diseases. Many cattle skin diseases are very dangerous, mostly if not treated at their first stage. These diseases have imposed economic losses in different aspects especially in day-to-day usage they reduce milk yield, leather quality, and performance in draft cattle. In this regard, the study and investigation of these diseases have been an interesting sector for computing experts. But there is only one paper that has been done using KNN. However, the methods they have used are inefficient. I.e. KNN is a lazy learner Algorithm and always computes the distance metrics due to this it is not applicable for diseases with high correlation. This paper was the second study on four common cattle skin diseases including lumpy, dermatophytosis, dermatophylosis, and wart diseases. In this work, we have developeda cattle skin disease identification model usingthe comparison of three filtering techniques (median, Gaussian, and Gabor filter), threshold segmentation, hybrid feature extraction (CNN, and HOG), and SVM for classifications. We Have to use the HOG technique for the fact that CNN was not invariant for image transformation (rotation and illumination change) while HOG was invariant in image transformation. Anaconda python programming tools have been utilized to complete the overall coding mechanisms. The images were sourced from Bahir Dar zuria woreda veterinary clinic and Bahir Dar university agriculture and environmental science campus using a Huawei Y635 model smartphone memory size 8GB and resolution of 840x854. We have collected a total of 765 original images for four classes and CNN needs a large dataset due to this we used augmentation techniques. The total number of augmented images is 2000 images. The images are in a jpg format. We have achieved 96.5% accuracy in CNN, 93% with HOG, and 98.75% using a hybrid feature.