【Objective】 To evaluate the relationship between the pericoronary fat attenuation index (FAI) and the image reconstruction parameters of computed tomography (CT) coronary angiography, including reconstruction kernel, iterative reconstruction algorithm and image thickness. 【Methods】 Forty-four CT coronary angiography scans were prospectively enrolled. All scans were reconstructed by three means as follows: ① Four different kernels (Soft_AA, Soft_BA, Soft_CA, and Soft_DA, sharpness from low to high) as the iterative reconstruction algorithm (KARL5) and image thickness (0.5 mm) remained unchanged. ② Filtered back projection (FBP) and iterative reconstruction kernel (KARL5) as the kernel and image thickness (0.5 mm) remained unchanged. ③ Different image thickness (0.5 mm and 1 mm) as the kernel (Soft_AA) and iterative reconstruction algorithm (KARL5) remained unchanged. The FAI of left anterior descending artery (LAD), left circumflex artery (LCX), and right coronary artery (RCA) was calculated using a dedicated software. Paired t-test and analysis of variance were used for statistical analysis. 【Results】 For LAD, LCX and RCA: ① The differences of FAI among different reconstruction kernels reached statistical significance (P<0.001), and FAI decreased as the sharper kernel was used. ② Compared with FBP, the FAI of KARL5-reconstructed images significantly increased (P<0.001). ③ Compared with 0.5 mm, the FAI of images with 1.0 mm thickness significantly decreased (P<0.001). 【Conclusion】 The kernel, iterative reconstruction algorithms, and image thickness all have a significant impact on the FAI of each coronary artery. When using FAI for clinical diagnosis, the effect of CT reconstruction parameters should be taken into account.