Background: Pain and blood loss after total knee arthroplasty (TKA) are unsolved clinical problems. Some studies reported that periarticular cocktail injection can effectively reduce pain and blood loss. However, there was no gold standard about the cocktail ingredient and injection location. More osteotomy and less soft tissue release in TKA with mild deformity; besides, plenty of nerves and blood vessels are contained in the periosteums and bone marrow. In this study, we aimed to detect the clinical results of subperiosteal cocktail application in TKA. Methods: Two groups were included according to the different injection location in our study. In group 1, cocktails were injected into the muscles, tendons, suprapatellar bursa, and subpatellar bursa surrounding knee joint. In group 2, cocktail injection was performed under the periosteum of the distal femur and proximal tibia. Our primary outcomes were visual analogue scale (VAS) and hemoglobin (Hb), and the secondary outcomes were wound healing, infection, deep vein thrombosis (DVT), operation time, and hospitalization. Results: At the first operative day, the mean (standard deviation) VAS score in a state of static was lower in group 2 compared with group 1 (0.98 ± 0.27 in group 1 and 0.86 ± 0.60 in group 2, p < 0.05). In the state of flexion and extension, the mean (standard deviation) VAS was 1.61 ± 0.66 in group 1 and 1.10 ± 0.57 in group 2 (p < 0.05). The mean (standard deviation) blood loss was higher in group 1 than in group 2 at the first postoperative day (440.19 (167.68) ml in group 1 and 333.67 (205.99) ml in group 2, p < 0.05). At the third day after surgery, the mean (standard deviation) blood loss was 686.44 (140.29) ml in group 1 and 609.19 (260.30) ml in group 2, and there was significant difference between these two groups (p < 0.05). Conclusions: We concluded that subperiosteal cocktail injection can significantly reduce pain and blood loss compared with periarticular cocktail injection after TKA.