Abstract: Three one-and half year old Gansu high mountain fine wool wethers (average live weight 22.5 kg) per treatment were fitted with permanent rumen fistula, and used in a 3×3 Latin square design to study the effects on rumen metabolic profiles of three different diet combinations: Control, 8% dried beet pulp supplement, and 8% apple pomace supplement. Each of the three diet circulatory periods included 10 d for preliminary testing followed by 9 d for sampling during the whole experiment. The average pH value of rumen liquid was similar in all treatments (P＞0.05). The propionic molar ratio of rumen liquid of sheep fed with diet Ⅱ tended to be lower (P=0.109), and the ratio between acetic and propionic was higher for Ⅱ than for Ⅰ (P＜0.05) and Ⅲ (P＞0.05), but the rumen fermentation in all treatments belonged to the acetic acid type. The time changes of TVFA for diet Ⅱ had a similar parabola, and the highest values of TVFA for diets Ⅰ and Ⅲ were measured at 3 or 5 h after feeding respectively. There were no significant differences between the three treatments (P＞0.05) in the means of total N, protein N, NH₃-N, urea N, and the concentrations of bacterial N in rumen fluid. For diet Ⅱ, the concentrates of total N were always higher with two peaks observed at 3 and 5 h after feeding, but the NH₃-N levels were lower and the peak appeared 1 h after feeding. All the highest values of total N for Ⅰ and Ⅲ were measured at 7 h after feeding, and the peak of NH₃-N was observed at 1 and at 7 h after feeding for Ⅲ, but not for I. It is suggested that there were no significant differences in the types of rumen fermentation and in the efficiencies of nitrogen utilization among the three treatments but diet Ⅱ showed some superiority, for stabilizing rumen surroundings, maintaining fibrolytic activity and utilizing N compared with Ⅰ and Ⅲ.