An experiment was conducted to validate urinary creatinine as a marker for urine output and optimize spot urine sampling in dairy cows. Total 12 lactating cows were used in a randomized block design. Cows were grouped and randomly assigned to 2 diets: a diet with corn silage (49% of DM) or a diet with alfalfa silage (48% of DM) with supplemental potassium (3.0% K in dietary DM). The experiment lasted for 21 d and total collection of urine (TC) was conducted for the last 3 d. Twelve spot samples of urine were also collected to represent every 2 h in a 24-h cycle during TC. Urine from total collection and individual spot samples from cows were analyzed for creatinine concentrations. Creatinine concentrations from individual spot samples were averaged to simulate equally spaced 12, 6, 4, and 2 time points of composite urine (12TPs, 6TPs, 4TPs, and 2TPs, respectively; starting at feeding). The coefficient of 29 (creatinine excretion, mg/kg of BW d⁻¹) was used to estimate urine output from TPs. All data were analyzed using the MIXED procedure of SAS with group as random and diets, methods, and their interaction as fixed. Urinary output of cows fed the diet with alfalfa silage and potassium increased (38.5 vs. 26.8 kg/d; P < 0.01) compared with the corn silage diet. Creatinine concentrations from TC varied from 36 to 121 mg/dL and coefficient from 16.7 to 34.5 with an average of 28.5. Creatinine concentration from 12TPs was greater (75 vs. 65 mg/dL; P = 0.05) when compared with TC, resulting in underestimating (P = 0.025) urine outputs. When compared among TPs, creatinine concentrations and urine outputs were different (P ≤ 0.09) between 12TPs and 2TPs but not different among 12, 6, and 4TPs. In the current study, creatinine concentrations from composite spot samples (even 12TPs) failed to represent true creatinine concentration (TC), resulting in inaccurate estimation of urine output. This occurred due to the large diurnal variation of urine creatinine: urinary creatinine concentration largely decreased after feeding and gradually increased 6 h after feeding for both diets.