Purpose: Emergency laparotomy is associated with significant morbidity and mortality, and often requires post-operative ICU transfer. While risk prediction models exist for mortality and morbidity, there is no validated model for post-operative ICU transfer. This study aimed to develop a predictive model to identify those at risk of post-operative ICU transfer and mortality following emergency laparotomy. Methodology: Prospectively collected demographic and clinical data from the Australian and New Zealand Emergency Laparotomy Audit – Quality Improvement (ANZELA-QI) was used to develop three predictive models: Logistic Regression (LR), Random Forest and XGBoost. These utilised selected predictor variables against the outcomes of mortality and ICU transfer. Results: Between July 2018 and July 2023, 8615 cases were collected from 35 Australian hospitals. From this, 5195 cases with a complete dataset were used to develop the predictive models. Mortality and post-operative ICU admission rates were 12% and 42%, respectively. For both outcomes, LR demonstrated greatest overall accuracy. For post-operative ICU transfer, sensitivity was 0.71 and specificity 0.76. For the death outcome overall accuracy was less, with sensitivity of 0.23 and specificity of 0.98. Conclusion: These models provided promising results for ICU prediction, but were far less accurate for mortality. Prediction of ICU transfer would provide an important adjunct in peri-operative risk assessment, facilitating earlier discussions with intensivists, improved triaging and better shared decision making with patients. Ongoing data collection, optimising predictor variables, model selection, and sub-grouping laparotomy by primary procedure may improve accuracy.