Tissue-directed stretching interventions can preferentially load muscular or non-muscular structures such as peripheral nerves. How these tissues adapt mechanically to long-term stretching is poorly understood. This randomized, single-blind, controlled study used ultrasonography and dynamometry to compare the effects of 12-week nerve-directed and muscle-directed stretching programs versus control on: maximal ankle dorsiflexion range of motion (ROM) and passive torque, shear wave velocity (SWV; an index of stiffness) and architecture of triceps surae and sciatic nerve. Sixty healthy adults were randomized to receive nerve-directed, muscle-directed stretching, or no intervention (control). The muscle-directed protocol was designed to primarily stretch the plantar flexor muscle group, while the nerve-directed intervention targeted the sciatic nerve tract. Compared with the control group (mean; 95% Confidence Interval), muscle-directed intervention showed increased ROM (+7.3°; 95% CI: 4.1-10.5), decreased SWV of triceps surae (varied from -0.8 to -2.3m/s across muscles), decreased passive torque (-6.8N.m; 95% CI: -11.9 to -1.7), and greater gastrocnemius medialis fascicle length (+0.4cm; 95% CI: 0.1 to 0.8). Muscle-directed intervention did not affect the SWV and size of sciatic nerve. Participants in nerve-directed group showed a significant increase in ROM (+9.9°; 95% CI: 6.2 to 13.6) and a significant decrease in sciatic nerve SWV (> -1.8m/s across nerve regions) compared with the control group. Nerve-directed intervention had no effect on the main outcomes at muscle and joint levels. These findings provide new insights into the long-term mechanical effects of stretching interventions, and have relevance to clinical conditions where change in mechanical properties has occurred.