Sugarcane is an option for feeding dairy cattle in the tropics. Reductions in DMI and milk yield (MY) are frequently reported when sugarcane replaces corn silage in dairy rations, due to the low NDF digestibility. This meta-analysis was conducted to evaluate DMI and MY response of dairy cows to the replacement of corn silage with sugarcane. A database was compiled from 13 Brazilian publications involving 50 treatments from the last 12 years. The nutrient composition of the diets was calculated using CNCPS v. 6.5 feed library. Treatments were stratified in 3 categories: no replacement (SC0), 25–75% replacement (SCA) and 100% replacement (SCB) and the outcomes and predictions were analyzed with a mixed effects model (random: study; fixed: replacement level). Then, a mixed model regression analysis was used to investigate the potential relationships among reported nutrient composition of the diets (CP, NPN, NDF, NFC, Lignin and CNCPS estimated fractions) and MY and DMI as dependent variables. The best models were selected based on the highest BLUP (R²_BLUP) correlation coefficient and the lowest root mean square error (RMSE). Total replacement reduced MY (21.9 vs. 18.6 kg/d) and DMI (19.1 vs. 16 kg/d). DMI was the single most important factor explaining MY (R²_BLUP = 0.973; RMSE = 0.865. When CP and RUP (%CP) were included into the DMI model, accuracy and precision were increased (R²_BLUP = 0.979; RMSE = 0.75. Total NDF of the diet did not affect MY, but sugarcane NDF content affected it negatively. Sugar (CHOA4; R²_BLUP = 0.948; RMSE = 0.66) and sugarcane peNDF (R²_BLUP = 0.941; RMSE = 0.68) were the 2 primary factors negatively related to DMI. The replacement of corn silage with sugarcane reduced MY and DMI of dairy cows, diet sugarcane NDF content was the nutrient most related to the response.