The soluble component of the feed protein (solCP on a CP basis) represents a valuable source of nitrogen for the microbial protein synthesis in the rumen. However, in condition of inadequate energy availability or high rumen protein degradation rate, ammoniogenesis in the rumen might result with lower nitrogen efficiency. Because of their ability of forming insoluble complexes with protein fractions, tannins might increase the by-pass of protein, reduce ammonia yield in the rumen and urea excretion in milk. Six multiparous cows (43±9 kg/d milk yield and 93±31 d in milk) were randomly assigned to 3 diet groups in a replicated Latin square design. Treatment diets (TRT) were CTR+ (14.9% CP on a DM basis, 28.2% solCP), CTR- (14% CP/DM, 31% solCP) and the T in which the CTR- diet was added 80 g/head/d of a chestnut hydrolyzable tannin. Each experimental period lasted 3 weeks. Cows were kept in a common pen, individually fed by means of a Calan Broadbent© feeding system and individual milk yield was recorded at each milking. Data measured over time were subjected to ANOVA using the repeated statement in the mixed procedure of SAS. The experimental unit was the animal treated alike. The statistical model included the fixed effect of TRT, time of measurement and (TRTxtime of measurement) interaction. The random variable was the animal within the square. The T group had a higher (P<0.05) DM intake (26.9 vs 25.6 kg/head/d) and lower (P<0.05) milk urea content (21.8 vs 23.8 mg/100ml) compared to the CTR+. No differences among groups were observed on milk (38.9, 37.5, 37.6 kg/head/d, geometric mean), fat (1.49, 1.42, 1.47 kg/head/d, geometric mean) and protein yield (1.33, 1.31, 1.31 kg/head/d, geometric mean), respectively for CTR+, CTR- and T groups. When adding chestnut tannin to high solCP diets, the urea excretion could be reduced as a result of a probable different protein degradation rate in rumen.