Since manual body condition scoring has been widely utilized as an indirect and subjective method to estimate energy reserves of dairy cattle, image analysis has been increasingly researched for use on large farms as an objective and effective measuring instrument for the estimation of body condition score (BCS) and body weight (BW). Recent advances in the technological development of the three-dimensional (3D) cameras may provide innovative feed management tools for dairy farms. The objective of the present study was to evaluate the feasibility of a 3D camera systems in measuring the back posture of lactating Holstein dairy cows to predict the BCS, BW, milk yield (MY), milk fat (MF) and milk protein (MP). The BCSs for eight cows were recorded by two trained observers using a 5-point scale, and other variables were obtained using an automatic milking system during the lactation. Back posture measurements of dairy cows were conducted using the ASUS Xtion Pro sensor. Six geodesic line (GL) lengths were computed using the 3D objects of each cow based on the positions of the right and left hook bones (GLhh), right and left thurl bones (GLtt), right and left pin bones (GLpp), hook and thurl bones (GLht), hook and pin bones (GLhp), and coccygeal ligament (GLcl). In the principal component analysis (PCA), GL, GLpp, and GLcl had the greatest contribution to principal component values (PCV) 1, 2, and 3, respectively, and these three PCVs described 0.887 of the cumulative contribution ratio. Good correlations were found between the observed and predicted values of BCS (R2=0.74), BW (0.80), MY (0.62), MF (0.62), and MP (0.53) based on linear regression equations using the GLs as explanatory variables and parity (1, 2, and >3) as a fixed effect. These results demonstrate that the 3D cameras could represent an innovative tool for estimating body condition and milk properties.