Background: the prevalence of esophageal varices (OV) in newly diagnosed cirrhotic patients is approximately 60-80% and the 1-year rate of first variceal bleeding is approximately 5% for small esophageal varices & 15% for large esophageal varices. Non invasiveness has become a major goal in hepatology in the latter years, several serum markers and imaging methods have been tried for the non invasive assessment of portal hypertension or presence of esophageal varices. Aim of the work: this study aimed to compare ALBI, MELD and Child-pugh scores in prediction of esophageal varices and for discrimination between risky and non risky esophageal varices. Methods: in this Prospective study evaluation of of ALBI, MELD And Child-Pugh Scores As non-Invasive Predictors of Esophageal Varices was done in 80 patients with liver cirrhosis. They were divided into 2 groups, Group I: included 60 patients with liver cirrhosis & esophageal varices diagnosed by upper GIT endoscopy and will be divided into 3 subgroups 20 patients each as described below: Small, Moderate and Large esophageal varices. Group II: included 20 patients with liver cirrhosis with no esophageal varices as the control groupg. Results: the current study showed that ALBI score could be used as a non invasive predictor of esophageal varices with a cut-off value > - 2.2, with 96.7% sensitivity, 100% specificity, Child score could be used with a cut-off value > 5.5, with 93.3% sensitivity, 100% specificity, and MELD score could be used with a cut-off value > 8.5, with 90% sensitivity, 95% specificity. Conclusion: ALBI score is more accurate than Child and MELD scores as non invasive predictor of esophageal varices and its grading