Background Increased serum level of lipoprotein(a)(Lp(a)) is associated with atherosclerosis. Whether increased Lp(a) level is independently associated with the severity of coronary artery disease(CAD) is unclear. Methods Subjects were enrolled and received coronary angiography to assess the number of stenosed coronary artery. The subjects with CAD were divided into non-significant(< 50% stenosis), single and multivessel stenosis(≥ 50 % stenosis) groups. Parameters of interest at baseline were collected. Statistical analyses were performed to evaluate the relationship between Lp(a) level and CAD severity. Results Totally 745 populations were enrolled and diagnosed as CAD(n = 605) or without CAD(n = 140) on the basis of angiography examination. As compared to the subjects without CAD, serum levels of Lp(a) and CRP, and the percentages of subjects with smoking or diabetes were significantly higher in subjects with CAD. In contrast,serum levels of HDL-C and Apo-A were significantly lower in subjects with CAD as compared to subjects without CAD. In comparison of subjects with non-significant stenosis(serum Lp(a) level, 170.0 ± 19.7 mg /d L), serum Lp(a) level was significantly higher in subjects with single(245.5 ± 22.3 mg / d L) or multiple vessel stenoses(265.8 ± 14.0 mg / d L). With multivariate regression analyses, after adjusted for age, gender,smoking, family history and hypertension, there was still significant association between serum Lp(a) level and the number of coronary artery stenosis. After additional adjustment for diabetes, Hb A1 c, total cholesterol,LDL-C, Apo-A, uric acid and CRP, Lp(a) remained strongly associated with CAD severity. Conclusion Serum Lp(a) level was significantly associated with the severity of coronary artery stenosis, which may add the value on cardiovascular risk evaluation.