To identify the characterization of culprit lesions in acute coronary syndrome (ACS) compared with stable angina pectoris (SAP) by dual-source computed tomography (DSCT). 65 patients with ACS and 75 controls with SAP and a similar atherosclerotic risk profile were studied. Computed tomography (CT) coronary angiography was performed using a DSCT scanner before invasive catheterization. Using DSCT and quantitative coronary angiography (QCA), lesion characteristics [luminal cross-section area (L-CSA), vascular cross-section area (V-CSA), plaque area and degree of stenosis) were detected. Plaque types, mean and minimum CT density (Hounsfield Unit; HU), remodeling index, and presence of “spotty” calcifications were analyzed using DSCT. A good correlation was observed between DSCT and QCA for all lesion characteristics (P < 0.05). Culprit lesions in ACS had much larger V-CSA (20.5 ± 6.0 vs. 14.8 ± 4.8 mm²), plaque area (15.3 ± 5.0 vs. 11.1 ± 3.3 mm²) and remodeling index (1.3 ± 0.2 vs. 1.0 ± 0.4) than stable lesions in SAP (P < 0.05). The prevalence of non-calcified/calcified/mixed plaque was 30/0/35 in ACS versus 25/15/35 stable lesions in SAP (P < 0.01). The proportion of “spotty” calcified plaques was 21.5 % in culprit lesions (14 of 65) versus 1.3 % in SAP (1 of 75). The mean/minimum HU of culprit lesions was 88.6 ± 43.2/154.2 ± 98.7 in ACS versus 45.9 ± 34.7/98.2 ± 76.8 in SAP (both P < 0.01). DSCT is a feasible means of detecting coronary stenosis with good accuracy compared with QCA. Culprit lesions in ACS display a greater proportion of non-calcified material with lower CT attenuation, “spotty” calcifications and higher remodeling index compared with SAP lesions.