The severity of canine acute pancreatitis (AP) is highly variable. Clinically applicable severity scoring algorithms are limited. One-hundred-thirty-four dogs diagnosed with AP and hospitalized between 2009 and 2014 were selected for analysis to develop a severity scoring system. Eleven variables, i.e. age, hemoglobin, red blood cell count, albumin, total bilirubin, blood urea nitrogen, creatinine, phosphorus, potassium, presence of systemic inflammatory response syndrome and oliguria/anuria after rehydration (Day 2 to Day 4) were significantly abnormal and selected by logistic regression to predict the outcome. A scoring system was developed by incorporating weighting factors assigned to each quartile based on the odds ratio integer value. A predictive score is the sum of all weighting factors for each dog. Receiver-operating curve analyses were performed to assess the sensitivity, specificity and optimal cut-off point of the outcome predictions. The scores of 134 dogs after rehydration ranged from 11 to 78 points. The mean score was 3 2. 9 ± 2 0. 0 , and the median score was 23 points. The optimal cut-off point for outcome prediction of the scoring system was 35.5, with a sensitivity of 96% and a specificity of 89.3%. The mortality was 85.5% with a score of > 3 6 , whereas 3.8% had a score of ≤ 3 6. Dogs with a score < 2 6 all survived. The severity scoring system provides a reliable and clinically applicable method to assess disease severity in dogs with clinically AP at the time point after rehydration. [ABSTRACT FROM AUTHOR]