Early recognition is a critical health concern with chronic liver disease since it allows for better management and treatment. The present research presents an integrated strategy utilizing statistical feature extraction and machine learning approaches to predict chronic liver illness and identify tattoo-induced jaundice. First, we employ statistical techniques based on inference to extract pertinent information from patient data, such as outcomes of tests and clinical markers. We then use neural networks (ANN), RNN, Random Forest method, and Naïve Bayes to effectively identify periods of tattoo-associated jaundice and classify them based on the likelihood that they have chronic liver ailment. NaïveBayes makes categorical data handling easier, ANN uses layers of neural networks to catch intricate patterns, and RNN provides interpretability. The performance of these models is evaluated using the following metrics: F1 rating, recall, accuracy, and precision. Our study intends to improve prediction accuracy while offering helpful suggestions for early diagnosis and customized treatment plans. The combined strategy has the potential to improve healthcare outcomes via the use of machine learning and advanced data analytics methods.