Objective To explore the effects of internal fixation in surgical treatment of intertrochanteric fractures by proximal femoral locking compression plate (PFLCP) and proximal femoral nail anti-rotation (PFNA) on fracture morphological fea-tures. Methods 43 patients with femoral intertrochanteric fractures were treated with PFLCP or PFNA from January 2013 to De-cember 2015. There were 23 patients treated with PFLCP and 20 cases with PFNA. There were 29 males and 14 females, the aver-age age was 62.6 years old (ranged from 46-85 years). Observed the fracture type (AO/OTA, Evans, the lateral trochanteric wall in-jury), the direction of key fracture line (lateral superior to medial inferior/medial superior to lateral inferior), preoperative days, pre-operative bone mineral density value, operation time, the volume of intraoperative blood loss, the supportive contact rate at the me-dial cortex of proximal femur after surgery, the surgical intervention rate for medial cortical fracture, postoperative partial load bearing time, full weight bearing time, complications of internal fixation and Harris scores. And analyzed the relationship between the direction of key fracture line and the internal fixation failure. SPSS13.0 software was used for statistical analysis. Results No statistical significance was found in age distribution, fracture types, preoperative days, preoperative bone mineral density value, the internal fixation failure rate (loss of fracture reduction, screw cutting or plate fracture). The operative time of PFLCP group was obviously longer than that of PFNA group and the difference was statistically significant (t=2.216, P=0.032). The intraoperative blood loss was more in PFLCP group than PFNA group and significant difference was found between the two groups(t=4.142, P=0.000).Though the initial postoperative partial load bearing time is earlier in the PFNA group than PFLCP group, but no statistical difference bewtween the two groups (t=0.833, P=0.902). Full weight bearing time had no statistical difference between the two groups. Harris scores in the PFNA group was slightly lower than in the PFLCP group, but there was no statistical difference ( t=0.833, P=0.902). The postoperative medial cortical support rate was higher in the PFLCP group than PFNA group, but no statisti-cal difference between them. The surgical intervention rate for medial cortical fracture was higher in the PFLCP group and differ-ence was statistically significant (χ2=4.768, P=0.029). No correlation had found between invalid support at medial cortex and the direction of key fracture line in all internal fixation failure cases. Distribution of the direction of key fracture line in the two groups had no statistical difference, while the relative risk ratio in terms of internal fixation failure between 2 types of the direction of key fracture line (OR=lateral superior to medial inferior/ medial superior to lateral inferior) had statistical difference (χ2=6.081, P=0.014;OR=9.600, P=0.037). Conclusion The direction of key fracture line can foresee the underlying displacement direction of intertrochanteric fractures at postsurgical load bearing, combining with restore the supportive contact at the medial cortex of the proximal femur and timing of load bearing may be determined to avoid internal fixation failure.