BACKGROUND Accurate prediction of the severity of acute pancreatitis will help in identifying patients at increased risk for morbidity and mortality. We wanted to evaluate the different scoring systems in predicting the severity of acute pancreatitis. METHODS This cross-sectional study was undertaken in the Department of Surgery at a zonal hospital between April 2013 and December 2014. RESULTS 40 patients were selected and enrolled in the study as per the selection criteria. 20 (50 %) patients had fair outcome and 20 (50 %) had a poor outcome. Accuracy of different scoring systems in predicting patient outcome ranged from 45 % (48-hr APACHE II) to 62.5 % (Goris MOF at baseline and 48 hr). Baseline Goris MOF was 70 % sensitive and 55 % specific in prediction of poor outcome. It had an accuracy of 62.5 % in prediction of outcome. 48-hr Goris MOF was 80 % sensitive and 45 % specific in predicting the outcome. Baseline APACHE II scores were below the cut-off level in all the patients. 48-hr APACHE II scores were 5 % sensitive and 100% specific for prediction of outcome. Ranson score > 3 was 25 % sensitive and 90 % specific in the prediction of outcome. Balthazar score > 6 was 65 % sensitive and 55 % specific in prediction of outcome. Ranson score was found to have a limited sensitivity for different outcomes (ranging from 21.1 % to 50 %) but was found to have a high specificity (83.8 % to 90 %). CONCLUSIONS Goris scoring system (at 48 hrs) was found to be highly sensitive to different poor outcomes as well as duration of hospital stay. It also correlated with Balthazar scoring system, which was also highly sensitive to different poor outcomes studied. KEYWORDS Acute Pancreatitis, Prediction, Scoring System, APACHE II, Goris MOF, Ranson’s Score, Balthazar Score