It is challenging to identify potential patients with acute coronary syndrome (ACS) in the emergency department (ED), although these cases should immediately undergo further evaluation in the observation unit. This study aimed to establish a new and rapid assessment system for triaging patients with potential ACS in the ED. Data from 1022 cases (June 2012–August 2015) were evaluated using latent class analysis to identify key symptoms and medical histories. Significant variables in the latent class analysis were entered as predictors for the new triaging system, and the final model was selected based on the false alarm rate, hit rate, and discriminability index. The new system provided better discriminability and significantly reduced the false alarm rate, compared to conventional methods. Our results indicate that symptom clustering analysis can facilitate the identification of potential ACS cases using a risk stratification system in the ED. The symptom clustering may facilitate a rapid assessment tool that reduces the costs of unnecessary diagnosis and hospitalization. Furthermore, this system might be developed as an application for embedding in ambient assisted living homes.