Objective: To compare the clinical effects of continuous negative-pressure wound therapy (NPWT) and conventional pressure dressing at at hard-to-fix sites after split-thickness skin grafting. Methods: From September 2017 to August 2019, 129 patients who met the inclusion criteria and had spilt-thickness skin grafting at hard-to-fix sites were admitted to the First Affiliated Hospital of Air Force Medical University and included in this retrospective cohort study. The patients were divided into NPWT group (67 patients, 41 males and 26 females, aged (32±6) years) and conventional pressure dressing group (62 patients, 37 males and 25 females, aged (30±5) years) according to whether the hard-to-fix sites were applied with NPWT after spilt-thickness skin grafting. After debridement and spilt-thickness skin grafting at hard-to-fix sites in patients of 2 groups, the wounds of patients in conventional pressure dressing group were applied with conventional pressure bandaging after being filled with dry gauze; for the wounds of patients in NPWT group, the semi-permeable membrane was pasted and sealed for continuous negative pressure suction after filled with dry gauze and placed the drainage foam or drainage tube, with the negative pressure ranging from -16.6 to -9.9 kPa. The bandage was opened during the first dressing change on the 5th day after surgery in NPWT group and on the 7th day after surgery in conventional pressure dressing group. The skin graft surviving area and proportion, the area and proportion of hematoma, the incidence of common complications of skin graft were observed and calculated. The times of postoperative dressing change and the length of hospital stay were counted. Data were statistically analyzed with two independent sample t test, Cochran & Cox approximate t test, chi-square test, and Fisher's exact probability test. Results: (1) At the first dressing change, the skin graft surviving area of patients in NPWT group was (420±94) cm(2), which was significantly larger than (322±97) cm(2) in conventional pressure dressing group (t'=12.33, P<0.01); the skin graft surviving area proportion of patients in NPWT group was (97.0±2.3)%, which was significantly higher than (74.4±4.8)% in conventional pressure dressing group (t'=50.11, P<0.01). (2) At the first dressing change, the skin hematoma area of patients in conventional pressure dressing group was (31.7±10.1) cm(2), which was significantly larger than (3.2±0.7) cm(2) in NPWT group (t'=23.04, P<0.01); the skin hematoma area proportion of patients in conventional pressure dressing group was (7.3±2.3)%, which was significantly higher than (0.7±0.3)% in NPWT group (t'=76.21, P<0.01). (3) At the first dressing change, there was 1 case of skin movement and no case of skin graft edge tear in NPWT group with an incidence of 1.5% (1/67). In the conventional pressure dressing group, there were 4 cases of skin movement and 2 cases of skin graft edge tear with an incidence of 9.7% (6/62), P<0.05. The incidence of complication of skin graft of patients in NPWT group was significantly lower than that in conventional pressure dressing group (P<0.05). (4) The times of postoperative dressing change of patients in NPWT group was significantly less than that in conventional pressure dressing group (t=7.93, P<0.01). The postoperative length of hospital stay in NPWT group was significantly less than that in conventional pressure dressing group (t=11.71, P<0.01). Conclusions: Continuous NPWT can effectively promote wound healing, improve the survival rate of skin graft, reduce the incidence of complications after skin grafting, and shorten the length of hospital stay in split-thickness skin grafting at hard-to-fix sites.