The precision of the measurement of the angulation of coronary stents or lesions using coronary angiography (CA) and computed tomographic angiography (CTA) has not been established, and obtaining a rotating artery tree to measure angulation based on CTA is time-consuming. The aim of this study was to evaluate the utility of a new three-dimensional centerline method that we have developed for the measurement of coronary stent angulation based on CTA and to compare it with other conventional methods. We used the centerline method compacted by means of our new software, the conventional artery rotation method based on CTA and the simple CA method to measure the angulations of phantoms in vitro and stents implanted in patients. The precision and repetition of this new method was compared with those of the other two methods. The angulation values obtained from both the centerline and artery rotation methods based on CTA had high correlation and agreement with the true angulation values measured using a phantom; the 95 % confidence intervals (CIs) for the differences were −0.67° to 0.91° and −0.59° to 2.93°, respectively, while the difference between the value determined using the CA method and the true angulation of the phantoms ranged from 3° to 21.8° (median 8.1°). In clinical coronary stent measurement, the difference between artery rotation and centerline measurement was small (95 % CI −9.0° to 7.6°), and both methods had good repeatability. The time required to complete the measurement was considerably shorter (p < 0.001) using the centerline method than artery rotation method (12.5 ± 1.86 vs. 71.8 ± 13.6 s), while the CA method had poor precision and repeatability in the measurement of clinical stent angulation relative to the methods based on CTA (95 % CI −14.7° to 21.7°). Our three-dimensional centerline method based on CTA for the measurement of angulation was reliable and easy to implement in both clinical and basic research image analysis, and the centerline and conventional artery rotation methods can be used interchangeably. In addition, the value obtained for the coronary stent angulation using the CA method had a large bias.