Background Some small prospective studies have shown that vitamin D deficiency was associated with the occurrence of diabetic peripheral neuropathy. To our knowledge, no previous study has assessed the effectiveness of vitamin D 2 for diabetic peripheral neuropathy. We hypothesised that in patients with diabetic peripheral neuropathy, vitamin D 2 supplements would improve vitamin D deficiency, clinical symptoms, and nerve conduction velocity. Methods We did a multicentre, randomised, double-blind trial that enrolled adults with type 2 diabetes with diabetic peripheral neuropathy diagnosed with WHO criteria (1999). Inclusion criteria were: patients aged 35–65 years with diabetic peripheral neuropathy for 5–10 years. Diabetic peripheral neuropathy diagnostic criteria was one of these three conditions: the ends of the limbs showed paresthesia or disorder, the total Toronto symptom score was more than 7·0, or single neuropathy conduct velocity of less than 45 m/s in the motor nerves or 40 m/s in the sensory nerves. Exclusion criteria were other diseases that lead to peripheral neuropathy and serious disease of the heart, liver, kidneys, and thyroid and parathyroid diseases. Participants' vitamin D concentrations were not known at baseline. Individuals were randomly assigned by random number table to receive a vitamin D 2 injection (observation group) or placebo (control group) for 3 months. Both groups received basic treatment such as controlling blood glucose and trophic nerve drugs. Patients in the observation group received a intramuscular vitamin D 2 injection a dose of 5000 IU/kg based on bodyweight, while the control group had a placebo injection of 0·9% sodium chloride once every 2 weeks over a period of 3 months. The clinical endpoint was a year after treatment initiation. Symptom effectiveness could be proved when one or more symptoms improved: the total score of Toronto symptom score reduced by 2 scores or more; the average limb nerve conduction velocity increased, including the sensory and motor nerve conduction velocity of the median nerves, ulnar nerves, sural nerves, posterior tibial nerves, and common peroneal nerves; the serum concentrations of 25-hydroxyvitamin D (25[OH]D); and the proportion of adverse reactions and side-effects. Outcomes were the effective rate, serum 25(OH)D, average limb nerve conduction velocity, and adverse events. The measurement data were analysed by single variance and by the t test, and the enumeration data were analysed by χ 2 . Ethical approval was granted by the hospital ethics committee of the Wan Medical College. Result Between Oct 12, 201, and Oct 12, 2015, 320 patients were included in the analysis either assigned to the control group (n=166) or the observation group (n=154), with no statistics difference in age and disease course at baseline. In comparison between the control group and the observation group, the effective rates were 62% (103 of 166) versus 75% (116 of 154) after 3 months treatment (χ 2 =59·86, p=0·0001) and 38% (63 of 166) versus 65% (100 of 154) at 1 year follow-up (χ 2 =23·28, p=0·0001). The concentrations of serum 25(OH)D at baseline were 45·9 nmol/L in the control group versus 44·6 nmol/L in the observation group (t=0·50, p=0·60), 45·6 nmol/L versus 59·7 nmol/L after 3 months treatment (t=6·89, p=0·001), and 44·7 nmol/L versus 55·5 nmol/L at 1 year follow-up (t=5·68, p=0·001). The proportion of patients with 25(OH)D deficiency was 101 (61%) of 166 in the control group versus 94 (61%) of 154 in the observation group at baseline (χ 2 =0·001, p=0·95), 100 (60%) of 166 versus 26 (17%) of 154 after 3 months treatment (χ 2 =62·91, p=0·0001), and 96 (58%) of 166 versus 52 (34%) of 154 at 1 year follow-up (χ 2 =18·61, p=0·0001). Average limb nerve conduction velocities were 35·9 m/s in the control group versus 34·6 m/s in the observation group at baseline (t=0·85, p=0·76), 55·6 m/s versus 59·7 m/s after 3 months treatment (t=3·29, p=0·55), and 44·7 m/s versus 57·6 m/s at 1 year follow-up (t=7·89, p=0·0001). Six (4%) of 154 patients had nausea, pruritus, ostealgia, metallic taste in the mouth, and other side-effects, there were no other serious adverse effects in the observation group and no side-effects in the control group. Interpretation Supplementation for 3 months with vitamin D 2 increased serum concentrations of 25(OH)D and improved nerve conduction velocity. Both clinical symptoms and average limb nerve conduction velocity were better after 1 year of follow-up in the observation group than that in the placebo group. Vitamin D 2 is effective and safe for treatment of diabetic peripheral neuropathy. Funding Education Commission of Henan Province, China.