To detect the influence of raloxifene (RLX) on fracture heal ing in rabbit. Methods Eighty healthy New Zealand white rabbits (44 females and 36 males) weighing 1.9-2.1 kg were used. A 0.5-cm bone defect model in the mid-diaphysis of the left forel imb radius was establ ished in 72 rabbits, which thereafter were divided into 4 groups (n=18 per group, 10 females and 8 males): groups A, B and C received 7.5, 15.0 and 30.0 mg/ (kg• d) RLX, respectively, from the 2nd to the 50th postoperative day; group D received no further treatment. The rest untreated 8 rabbits (4 females and 4 males) served as normal control for serum osteocalcin detection. At different postoperative time points, bone mineral density detection, X-ray scanning, biomechanics measurement, histology and immunohistochemistry observations were conducted; serum estradiol, plasma cholesterol, serum osteocalcin and the ratio of uterine weight to body weight were detected. Results The bone mineral density of each group reached a peak 20 days after operation, showing a significant difference between groups A, B and C and group D (P lt; 0.05), and no significant differences among groups A, B and C (P gt; 0.05). On the 30th and 50th postoperative day, the maximum failure load and the maximum displacement of groups A, B andC were greater than those of group D (P lt; 0.05), but no significant differences among groups A, B and C were evident (P gt;0.05). On the 7th, 20th and 30th postoperative day, the X-ray score of fracture heal ing of groups A, B and C was greater than group D (P lt; 0.05); on the 50th postoperative day, there was significant difference between groups B and C and group D, and between group A and group C (Plt; 0.05), and no significant difference was evident between group B and group C (P gt; 0.05). The percentage of new bone formation in the fractured area of groups A, B and C was greater than that of group D on the 30th and 50th postoperative day (P lt; 0.05). For the type II collagen protein secretion in the fractured area, groups B and C were superior to group D on the 30th postoperative day (P lt; 0.05), and there was no significant difference between group A and group D (P gt; 0.05); no significant differences among four groups were evident on the 50th postoperative day (P gt; 0.05). On the 10th, 30th and 50th postoperative day, the serum osteocalcin of groups A, B, C and D was higher than that of normal control (P lt; 0.05), groups B and C were higher than group D (P lt; 0.05), and there was no significant difference between groups A, B and C, and between group A and group D (P gt; 0.05). For the plasma cholesterol, on the 30th postoperative day, no significant change was detected in each group (P gt; 0.05); on the 50th postoperative day, obvious decrease was observed in groups A, B and C, showing a significant difference compared with group D (P lt; 0.05). On the 30th and 50th postoperative day, there was significant difference between groups B and C and group D in serum estradiol (P lt; 0.05), and no significant differences were evident among other groups (P gt; 0.05). On the 30th and 50th postoperative day, the ratio of uterine weight to body weight in groups B and C was less than that of group D (P lt; 0.05), and no significant difference was evident between group A and group C (P gt; 0.05). Conclusion Oral administration of 7.5 mg/ (kg• d) RLX can promote the fracture heal ing of rabbit radius defect models safely and effectively.