Abstract Background and Aims The monitoring of vitamin D status is important to manage metabolic bone disease in patients with chronic kidney disease (CKD). 25(OH)D is used as the vitamin D marker in CKD patients, but vitamin D metabolite ratio (VMR) is also becoming useful as the marker, too. The classification of CKD is based on the cause-glomerular filtration rate (GFR)-albuminuria (CGA) in the KDIGO guidelines. This classification is importantly used in the management decision of CKD and predicts well the prognosis related to CKD. However, there are no study on the changes in various vitamin D markers according to CGA classification. We aimed to investigate the changes of vitamin D biomarkers according to classification by cause of CKD, estimated GFR (eGFR), and proteinuria of CKD patients. Method We prospectively analyzed blood and urine samples from a total of 206 patients who received informed consent with CKD class G2-G5. After classifying each group according to the presence or absence of diabetes, eGFR degree, and proteinuria amount, the differences in various vitamin D biomarkers in each group were compared. VMR was the ratio of 24,25(OH)2D to 25(OH)D. Results The mean age of the 206 patients was 64.14±12.72 years old. The patients with DM were 46.6% and the most common cause of CKD was glomerular disease (51.4%) including diabetic nephropathy (DN). There was no significant difference in all vitamin D markers we measured in the comparison between the DKD group and the non-DKD group. Among DKD patients, the DN group had significantly lower levels of 24,25(OH)2D (p=0.012) and bioavailable 25(OH)D (p=0.044) than the no DN group. When divided into three groups according to the degree of eGFR, the mean value of 24,25(OH)2D (p=0.003) and VMR (p<0.001) were significantly lower as the eGFR decreased but all 25(OH)D markers showed no significant decrease with the change in eGFR. In the diabetic patients, when divided into four groups according to the amount of proteinuria, the group with high proteinuria had significantly lower levels of total 25(OH)D (p=0.001), bioavailable 25(OH)D (p<0.001), free 25(OH)D (p=0.001), and 24,25(OH)2D (p=0.029) compared to the group with low proteinuria but there was no significant difference in VMR. In the non-diabetic patients, when divided into three groups according to the amount of proteinuria, the group with high proteinuria had significantly lower levels of total 25(OH)D (p=0.032), bioavailable 25(OH)D (p=0.010), and free 25(OH)D (p=0.035) compared to the group with low proteinuria but there was no significant difference in 24,25(OH)2D and VMR levels. In these CKD patients, there was no significant difference in the level of VDBP despite the presence or absence of diabetes, the degree of eGFR, and the amount of proteinuria. Conclusion As eGFR decreased, the levels of 24,25(OH)2D and VMR significantly decreased. All vitamin D markers we measured showed no significant difference depending on the presence or absence of diabetes except for the low 24,25(OH)2D and bioavailable 25(OH)D levels in DN patients. Regardless of the presence or absence of diabetes, all 25(OH)D markers decreased significantly as proteinuria increased. Although we showed that significant changes in vitamin D markers differed according to CGA classification, large-scale study and long-term follow-up are necessary for meaningful use in diagnosis and treatment.