ABSTRACT Objective Whole-plant corn silage (CS) is the main forage fed to dairy cows in China, usually comprising about 35% of the total DM of the TMR. However, sorghum is more resistant to drought and tolerates soils that are less fertile compared with corn, and consequently, sorghum silage (SS) would be preferable to CS in many drought-prone areas of China. Sweet SS is characterized by a high content of water-soluble carbohydrates and energy yield. Previous studies examined the effects of feeding either CS or SS in the TMR of lactating cows. The aim of the present study was to assess the effects of substituting different levels of sweet SS for CS in the TMR fed to dairy cows in a smallholder farm in western China. Materials and Methods Sixteen multiparous Holstein cows (mean BW of 454 ± 25 kg), producing approximately 12 kg/d milk, were assigned randomly to 4 dietary treatments. Experimental diets were formulated to be iso-energetic and iso-nitrogenous with 45% forage and 55% concentrate, and the diets included 10% oat hay and variable proportions of CS and SS. Dietary treatments included (1) 100% CS (35% DM of TMR, control diet); (2) 35% of CS replaced by SS; (3) 65% of CS replaced by SS; and (4) 100% of CS replaced by SS. Results and Discussion The DMI decreased linearly (P = 0.002) as SS increased. Milk yield and lactose percentage did not differ among treatments, but lactose yield exhibited a cubic response (P = 0.004), whereas milk fat percentage tended to increase linearly (P = 0.077) as SS increased. Increasing SS resulted in a quadratic response in milk protein percentage and yield (P = 0.001) and MUN increased linearly as SS increased (P = 0.022). The ECM/DMI ranged between 0.75 and 0.82 and increased linearly as SS increased (P = 0.029). Although there were differences among treatments in serum concentrations of aspartate transaminase, alanine aminotransferase, alkaline phosphatase, urea nitrogen, and glucose, all concentrations were in the normal range for healthy, lactating cows. Serum β-hydroxybutyric acid concentrations were well below the subclinical levels in all treatments. The SCC decreased linearly (P = 0.024) with increasing SS. Implications and Applications We concluded that for low-producing dairy cows, SS can replace CS in total in the diet without adverse effects on milk yield and its components and on serum metabolites. This is particularly important for small dairy herds in dry areas.