Diagnosing pediatric liver disease is difficult due to its unique physiological features and limited data accessibility. Artificial intelligence (AI) and machine learning (ML) have lately made significant advances in the development of non-invasive, efficient, and accurate assessments for the early detection, diagnosis, and treatment of pediatric liver disease. This study investigates and evaluates a wide range of techniques, including Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting (GB), and hybrid models, using a variety of datasets. Predictive values, sensitivity, specificity, and accuracy are measures used to assess the strengths and limitations of the current methods. Furthermore, to improve diagnosis accuracy and reliability, this study investigates the intriguing possibility of merging hybrid models with future data sources, such as imaging and biomarker data. Along with highlighting current research gaps and opportunities for further investigation, our analysis of these papers seeks to accelerate the development of specialized, efficient diagnostic tools for pediatric liver disease.