Pulmonary fibrosis (PF) is a respiratory disease with end-stage pathological changes of interstitial lung disease that severely affects the survival of patients. Among the many biomarkers that have been identified, serum Krebs von den Lungen-6 (KL-6) is by far the most frequent marker for detecting pulmonary fibrosis. We searched Medline (Pubmed), Embase, Web of science, and Cochrane databases for articles published between inception and August 2022 in order to explore the association between KL-6 and pulmonary fibrosis. Characteristics of patients and studies included in the articles were extracted by two independent investigators according to the inclusion and exclusion criteria. We reflected the accuracy of KL-6 in distinguishing between PF and non-PF by calculating sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and area under the curve by SROC curves. The presence of heterogeneity was reflected by I2 in the forest plot, and then the source of heterogeneity was investigated by meta-regression. We searched for 939 research papers, of which 16 met the inclusion criteria. Meta-analysis showed a sensitivity of 0.87 (95 % CI: 0.78–0.92), specificity of 0.91 (95 % CI: 0.86–0.95), positive likelihood ratio of 10.2 (95 % CI: 6.1, 17.0) and negative likelihood ratio of 0.14 (95 % CI: 0.08, 0.25) for KL-6 in diagnosing pulmonary fibrosis. The area under the receiver operating characteristic curve was 0.95 (95 % CI: 0.93–0.97). The results showed significant heterogeneity in both sensitivity and specificity (I2 = 94.55 and 91.52, respectively). Meta-regression analysis identified race as the cause of sensitivity heterogeneity and assay methodology as the cause of specificity heterogeneity. The analysis of this study suggests that serum KL-6 is a better tool for the diagnosis of pulmonary fibrosis when factors such as disease cause and control group category are not specifically considered.