In Mexico, dairies overfeed CP without improvements on milk production. Our objective was to assess the effects of 2 levels of CP and 2 ratios of rumen degradable:undegradable protein (RDP:RUP, respectively) on nitrogen (N) balance and milk production of dairy cows. Eight multiparous Holstein cows (640 ± 35 kg of BW and 124 ± 14 d in milk) were placed in a replicated 4 × 4 Latin square with 2 × 2 factorial arrangement of treatments to evaluate 4 diets: Low CP and high CP (LP and HP; 14.9 and 17.1% of dry matter, respectively) with RDP:RUP CP ratios of either 68:32 or 62:38. Levels of CP and RDP:RUP ratios were manipulated with canola meal and alfalfa hay. All diets contained 29.5% neutral detergent fiber, 25% starch, and 1.55 Mcal of NEl/kg. Cows were fed once and milked 3 times per day. On d 18, 19, and 20 of each 21-d period, milk samples, urine and fecal spot samples were taken at 6 h before and 6 h after feeding. Data were analyzed using proc Mixed of SAS. Dry matter intake (DMI) was greater for cows fed HP than LP diets (26.0 vs. 25.3 kg/d) and for cows fed 68:32 than 62:38 (25.7 vs. 25.1 kg/d) (both P < 0.05). Treatments did not influence milk yield (MY, 39.4 ± 1.8 kg/d), milk component (3.23 ± 0.02, 3.46 ± 0.05, and 5.0 ± 0.01% of protein, fat, and lactose, respectively), MY by component (1.2 ± 0.1, 1.3 ± 0.1, and 1.9 ± 0.2 kg/d of protein, fat, and lactose, respectively), and feed efficiency (MY/DMI; 1.5 ± 0.12). Cows fed HP had greater N intake (NI) than LP diets (688 vs. 602 g/d, P < 0.05), but no differences were detected for RDP:RUP ratios. Urinary-N, urinary urea-N, and milk urea-N were greater for cows fed HP than LP (227 vs. 148 g/d, 108 vs. 72 g/d, and 22.6 vs. 13.0 mg/dL respectively, all P < 0.05), but no effects of RDP:RUP ratios were detected. No differences were observed for fecal-N and milk-N (226 ± 8.4 and 195 ± 15.3 g/d, respectively), but N efficiency (milk-N/NI) was greater in cows fed LP than HP diets (34.3 vs. 28.8%, P < 0.05). In conclusion, LP diets did not penalize MY, but increased NE, and reduced N losses. However, RDP:RUP ratios had no influence on MY and N balance.