Phytophagous insects show feeding preference for different plant species, and such preference may be regulated by their demands for proteins and carbohydrates and the balance between them. Therefore, feeding preference of phytophagous insects may be closely related to carbon (C) content, nitrogen (N) content, and the C:N ratio of plants. Although many studies have investigated feeding preference of phytophagous insects for different plant species, relatively few have examined how plant functional groups affect feeding preference of phytophagous insects and whether their leaf C and N content, and C:N ratio matter. To examine feeding preference of phytophagous insects for different plant species and for different functional groups and its relationships with leaf C, N, and the C:N ratio of plants and functional groups, we conducted a feeding experiment in which we fed the larvae of a phytophagous insect, Gynaephora menyuanensis (Lymantriidae), with leaves of 31 common herbaceous plant species of four functional groups (sedges, grasses, legumes and forbs) collected in an alpine meadow dominated by Kobresia humilis on the Tibetan plateau. Fresh leaves of the 31 plant species were cut into pieces of the similar area (about 0.5 cm²) and fixed randomly on the arenas made of foam boards installed at the inner bottoms of six glass boxes (each measuring 1 m long × 1 m wide × 0.25 m high). Each box was treated as a block, containing ten leaf pieces of each of the 31 plant species collected on the same day. We put on the arena of each box 20 larvae of G. menyuanensis that were collected in the same meadow and had been starved for 24 hours, and measured consumption of each leaf piece (i.e., the percentage of leaf area loss) after 72 hours. Leaf area consumed by the larvae was further transformed into dry mass based on the relationship between leaf area and dry mass of each plant species. Of the 31 plant species, 19 were consumed by the larvae to different degrees, but the other 12 were not. The larvae showed strong feeding preference only for four species, i.e., K. humilis, Carex przewalskii, Helictotrichon tibeticum and Elymus nutans. At the species level, leaf consumption by the larvae was significantly negatively related with plant leaf N content and significantly positively related with the leaf C:N ratio. At the functional-group level, the sedges had the highest leaf consumption by the larvae, the lowest leaf N content and the highest leaf C:N ratio, whereas the legumes had the lowest leaf consumption, the highest leaf N, and the lowest leaf C:N ratio. We conclude that increasing N content of plants may not increase the feeding preference of phytophagous insects even in alpine meadow ecosystems where plant growth is highly limited due to the lack of soil available N. Our results also suggest that the strong feeding preference of G. menyuanensis larvae for the dominant species K. humilis and E. nutans may help in maintaining species coexistence and biodiversity in the alpine meadow dominated by K. humilis.