Objective To evaluate the biocompatibil ity of manufactured heterogeneous demineral ized bone matrix(DBM) particles and to provide basis for further experimental study and cl inical application. Methods Heterogeneous DBMparticles A (degreased and demineralized) and B (degreased, demineralized and acellular), particle size from 250 to 810 μm, and leaching l iquor were made with a series of physical and chemical methods from pig l imbs cortical bone. The residual calcium and phosphorus contents of bone particles were measured after degreased and demineral ized. The acute toxicity test, skin stimulating test, pyrogeneous test, hemolysis test, cellular toxicity test and muscular embedded test were carried out according standard toxicological method. Results The contents of calcium and phosphorus in cortical bone were (189.09 ± 3.12) mg/g and (124.73 ± 2.87) mg/g, and in demineral ized bone matrix particles were (3.48 ± 0.09) mg/g and (3.46 ± 0.07) mg/ g. The residual calcium content was 1.87%, of phosphorus was 2.69%. The activity of mice was normal in the acute toxicity test. No animal died and no toxicity symptom or adverse effects were shown within 7 days. The mean weight daily increased showed no statistically significant difference (P gt; 0.05) between two groups after 7 days. Skin stimulating reactions were not found in the two experimental groups and negative control group by intradermal stimulation test. The maximal increase of body temperature in two experimental groups were 0.4℃ , which meet the national standard (lt; 0.6 ). The rate of haemolysis to the leaching liquor was 1.14% (A) and 0.93% (B), which was lower than the national standard (lt; 5%). The cell prol iferation rates of two experimental groups when compared with control group showed no statistically significant difference (P gt; 0.05). The toxicity of DBM particlesleaching liquor was graded from 0 to 1, which means the material has no cytotoxicity. All the animals survived well. There was no tissue necrosis, effusion or inflammation at all implantation sites. For the index of HE and Masson staining, there were no effusion around the material and inflammatory cell infiltrate obviously in two experimental groups. Inflammatory cell infiltrate is sl ight in control group 2 weeks postoperatively. The inflammatory cell infiltration was mitigate gradually over time in two experimental groups after 4, 8 and 12 weeks. New bone and collagen fibers formation were observed when the material was degraded and absorpted. Score evaluation of local cellular immune response at different time after operation of two experimental groups showed no statistically significant difference (P gt; 0.05). Conclusion Heterogeneous DBM has no obvious toxicity, skin irritation, pyrogenicity, and no cytotoxicity with a rate of haemolysis lt; 5%, so it has good biocompatibility and partial osteoinductive.