To tackle the issue of disease outbreaks due to late detection of disease-infected chickens, this work aims to study the correlation of the chicken comb’s optical chromaticity with the disease infection in broiler chicken. Using images from online resources, the distinctive features of the chromaticity of the chicken comb were extracted in phase 1 of the project using the CIE XYZ color space to classify healthy chickens and chickens infected with various diseases using machine learning models (MLM). In phase 2, an experimental chicken trial was carried out for 50 days and the chickens were infected with the Newcastle Disease Virus (NDV). Chicken images were captured directly in the cages. Morphological and chromaticity features were extracted at 36 h post-infection (after NDV infection) with a 12-h time step onwards. Chromaticity analysis in phase 1 showed that the change of color from yellow to blue, as well as from red to green, occurs in the comb of chickens infected with diseases. The development of various MLM using the chromaticity features showed that disease-infected chickens can be classified with 95% accuracy. For Phase 2, it was established that within 36 h of NDV-infection, Linear Regression and Support Vector Machine models were able to detect infected chickens at a validation accuracy >85%. At 96 h post infection, > 95% testing accuracy and >87% validation accuracy were achieved. This is the first report on detecting NDV-infected chicken as early as 36 h post-infection using image processing and ML.