Objective To investigate the biomechanical stability of both DHS (dynamic hip screw) and PFN (proximal femoral nail) for treating unstable intertrochanteric fractures. Methods A standard 4-part osteotomy was performed in 8 pairs of fresh frozen human cadaver femurs, which were then randomly assigned to two groups: PFN group and DHS group for biomechanical testing. These specimens were applied to a cyclic load up to 200, 400, 600, 800, 1 000, 1 200, 1 400 N, respectively. Fracture displacement was measured during the loading to determine biomechanical stability of the implant. Each specimen was repeatedly loaded for 5 times to calculate the average displacement and draw the load-displacement curve. For failure testing, the initial load and loading rate was set at 1 400 N and 10 N/s, respectively. The applied compressive load was increased by 600 N each time for five cycles. The pressure was gradually increased to its peak force, and sustained for 10 second before it was gradually decreased to 0 N. The highest force value sustained before failure was defined as the maximum strength of the implant. Results The biomechanical testing on all specimens was completed successfully. There was no damage to the internal fixation. The average displacement and stiffness in DHS group were (3.92±2.21) mm and (215.28±58) N/mm, while those in PFN group were (4.22±1.80) mm and (197.06±34.20) N/mm, so no significant difference was found between the DHS and the PFN group (P> 0.05). New fracture occurred at the distal end of nail in PFN group. The DHS was fractured at the distal cortical screw, but no nail was cut out of the femoral head. The average load required for failure was (4 312±560) N in PFN group and (3 954±520) N in DHS group, and no significant difference was found between the two groups(P＞0.05). Conclusions The test shows that the PFN does not appear to offer any distinct biomechanical advantage over the DHS in the treatment of unstable intertrochanteric fractures. The implant chosen for treating intertrochanteric fractures must depend on patient’s fracture geometry, and anatomic reduction should be conducted in clinical treatment. If the anatomic reduction is difficult, trying to recover continuity of the posterior cortical bone would be necessary.