Background: Immediate risk stratification of patients with myocardial infarction in the emergency department (ED) at the time of initial presentation is important for their optimal emergency treatment. Current risk scores for predicting mortality following acute myocardial infarction (AMI) are potentially flawed, having been derived from clinical trials with highly selective patient enrolment and requiring data not readily available in the ED. These scores may not accurately represent the spectrum of patients in clinical practice and may lead to inappropriate decision making. Methods: This study cohort included 1212 consecutive patients with AMI who were admitted to the Mayo Clinic coronary care unit between 1988 and 2000. A risk score model was developed for predicting 30 day mortality using parameters available at initial hospital presentation in the ED. The model was developed on patients from the first era (training set—before 1997) and validated on patients in the second era (validation set—during or after 1997). Results: The risk score included age, sex, systolic blood pressure, admission serum creatinine, extent of ST segment depression, QRS duration, Killip class, and infarct location. The predictive ability of the model in the validation set was strong (c = 0.78). Conclusion: The Mayo risk score for 30 day mortality showed excellent predictive capacity in a population based cohort of patients with a wide range of risk profiles. The present results suggest that even amidst changing patient profiles, treatment, and disease definitions, the Mayo model is useful for 30 day risk assessment following AMI.