Objective To explore the effect of bone cement distribution, cement leakage, and clinical outcomes with side‐opening cannula for bone cement injection in percutaneous vertebroplasty (PVP) in treatment of Kummell disease. Methods A prospective study of patients with Kummell disease undergoing PVP was conducted from April 2012 to September 2017. In total, 43 patients (11 males, 32 females) with Kummell disease who received bilateral PVP were included in the study. The patients were divided into front‐opening cannulas (FOC) group with front‐opening cannulas and side‐opening cannulas (SOC) group with side‐opening cannulas. All patients were followed up for 6 months. The patient general information such as gender, age, bone density, compression ratio, operative time, and location of fracture vertebrae were recorded. Visual analogue scale (VAS), Oswestry Disability Index (ODI), bone cement distribution, radiation exposure time, bone cement leakage rate and vertebral height, and kyphosis angle were measured and compared for two groups before surgery, 1 day and 6 months after surgery. Results A total of 43 patients were enrolled, including 11 males and 32 females, aged 61–84 years. The bone density (T value) was 2.5 ± 0.6 in FOC group and 2.4 ± 0.6 in SOC group (P > 0.05). The compression ratio and operative time were 36.1% ± 13.0%, 39.3 ± 7.9 min in FOC group and 35.2% ± 13.7%, 40.0 ± 10.7 min in SOC group (P > 0.05). There was no significance between FOC and SOC groups in the location of fracture vertebrae. All patients underwent at least 6 months of follow‐up. At 6 months postoperatively, the VAS and ODI were significantly higher in the FOC group (3.0 ± 0.8, 35.7% ± 2.1%) than in the SOC group (1.3 ± 0.4, 18.6% ± 2.4%) (P < 0.05). The cement leakage rate of the SOC group was 4.8%, which was lower than that of the FOC group (31.8%, P < 0.05), and the bone cement distribution ratio was higher than that of the FOC group (63.1% ± 7.9% vs 40.5% ± 8.6%, P < 0.05). At 6 months after operation, the height of the anterior and posterior vertebral bodies of the patients in the SOC group restored better than the FOC group (anterior SOC: FOC 5.1 ± 0.5 mm vs 4.5 ± 0.5 mm; posterior SOC: FOC 0.6 ± 0.1 mm vs 0.3 ± 0.1 mm, P < 0.05), and the kyphosis correction was more obvious than patients in FOC group (SOC: FOC 8.5° ± 1.4° vs 4.6° ± 0.8°, P < 0.05). Conclusion Percutaneous vertebroplasty with side‐opening cannula is safe and effective in avoiding bone cement leakage, improving bone cement distribution, and restoring vertebral height.