Four early-lactation crossbred cows (82.5 % Holstein) were selected to investigate the effect of carbohydrate source and cottonseed meal level in the concentrate on rumen fermentation and milk production. Cows were randomly assigned to receive four dietary treatments according to a 2 × 2 factorial arrangement in a 4 × 4 Latin Square design. Factor A was carbohydrate source: cassava chip (CC) and CC + rice bran at a ratio 3:1 (CR3:1), and factor B was variation in the level of cottonseed meal (CM): low (LCM) and high (HCM) in isonitrogenous diets (180 g CP/kg DM). It was found that carbohydrate source did not affect feed intake, dry matter digestibility, rumen fermentation, microbial population, milk yield and composition, or economic return (P > 0.05). However, cows fed with CC had a higher population of amylolytic bacteria than cows fed with CR3:1 (P < 0.05). Cows fed with HCM had a higher total feed intake, milk yield and composition, and milk income when compared with cows fed on LCM although the concentrate and roughage intakes, dry matter digestibility, rumen fermentation, and microbial populations were similar between treatments (P > 0.05). In addition, the carbohydrate source and cottonseed meal level interactions were not significant for any parameter. It could be concluded that cassava chip and high level of cottonseed meal could usefully be incorporated into concentrates for dairy cows without impacting on rumen fermentation or milk production.