The purpose of this study is to evaluate the effect of extracorporeal shock wave therapy (ESWT) according to treatment timing in rabbits with 10% dextrose‐induced carpal tunnel syndrome (CTS); 0.1‐ml 10% dextrose solution was injected under ultrasound guidance twice weekly to the left forepaw subsynovial connective tissue (SSCT) within the carpal tunnel of 36 New Zealand white rabbits to induce CTS. The rabbits were randomly allocated into four groups: G1‐S (sham ESWT), G2‐E4 (ESWT at 4 weeks), G3‐E8 (at 8 weeks), and G4‐E16 (at 16 weeks). Radial ESWT (500 pulses, 0.08 mJ/mm2, 2 Hz) was repeated thrice weekly. Median nerve distal motor latency (DML) was measured before injection and at 4, 8, 12, 16, and 20 weeks after the first injection. All rabbits were sacrificed 20 weeks after injection. The median nerve cross‐sectional area (CSA) and SSCT thickness were measured with light microscopy. The mean median nerve DML at 4 weeks after the first dextrose injection did not differ from that at preinjection in all groups. The mean median nerve DML significantly increased before ESWT in all groups (p < .05); however, it did not increase in G2‐E4 and G3‐E8 for 12 weeks after ESWT and in G4‐E16 for 4 weeks (p > .05). Mean CSA of the median nerve and mean SSCT thickness in G2‐E4 were significantly lower than those in the other groups (p < .05). ESWT may prevent the progression of CTS for 12 weeks in rabbits with dextrose‐induced CTS regardless of treatment timing, and early application results in superior outcomes.