A 2 × 2 factorial arrangement of treatments was used to investigate the interactive effects of dietary macronutrient composition [high-starch, low-fat, low-fiber (HsLFF) than low-starch, high-fat, high-fiber (LsHFF)] and N source [soybean meal (SBM) or partially replaced by slow-release urea (SRU)] on lactation performance, rumen fermentation, N utilization efficiency, nutrient digestibility, blood metabolites, and feeding behavior in cows. A replicated 4 × 4 Latin square design was used involving 12 multiparous Holstein cows (milk yield of 40.5 ± 5.6 kg/d, BW of 590 ± 20 kg; 81 ± 12 d in milk). The HsLFF diet contained 300 g/kg starch, 31.3 g/kg fat, and 301 g/kg NDF, without straw or additional fat. In contrast, the LsHFF diet contained 195.5 g/kg starch, 60.8 g/kg fat, and 367.5 g/kg NDF, enriched with wheat straw (100 g/kg), and additional fat (34 g/kg). The diets were formulated to be iso-nitrogenous and isocaloric. Cows fed the HsLFF diet had greater DM intake, digestibility of DM and CP, milk yield and milk protein %, but lower intakes of NDF, and physically effective NDF, and milk fat % than cows fed the LsHFF diet. Replacing SBM with SRU significantly increased milk solids yield without affecting other lactation performance or BW. Cows fed the LsHFF diet had higher ruminal pH and branched-chain VFA proportions but lower total VFA concentrations compared with the HsLFF diet, while those on the LsHFF-SRU diet had the highest ruminal ammonia levels. Compared with the HsLFF diet, cows fed LsHFF had lower NE intake, milk energy output and energy requirement for maintenance, although energy balances were similar among groups. The HsLFF diet improved N utilization, resulting in higher N content in milk and lower N excretion in feces. Blood metabolite studies showed significant interactions between the main factors, particularly for blood glucose and creatinine, with the lowest levels in cows fed the LsHFF-SRU and LsHFF-SBM diets. In addition, alanine aminotransferase levels were higher in cows fed the LsHFF diet than in cows fed the HsLFF diet. This could indicate early-stage liver stress due to the metabolic imbalance caused by a high-fat, low-starch diet, which can alter energy metabolism. Cows on the HsLFF-SRU diet had the highest glucose levels, indicating possible changes in carbohydrate metabolism or a higher metabolic rate. The concentration of BUN increased steadily after feeding in the LsHFF diet and peaked after 4 h in the LsHFF-SRU diet, with no difference between N sources in the HsLFF diet. Partial replacement of SRU with SBM had no effect on BUN. The interaction between the main factors had a significant effect on MUN content, which was highest in LsHFF-SRU and lowest in HsLFF-SRU, with no difference between the N sources in the HsLFF diets. Overall, while diets with reduced starch and increased fiber and fat compromised lactation performance, partially substituting SRU with SBM helped maintain milk production and milk nitrogen efficiency. However, the LsHFF-SRU diet was less efficient in N utilization, as shown by higher levels of ruminal ammonia, BUN and MUN.