Objective To evaluate the relationship between the blood flow values generated from MR arterial spin labeling (ASL) perfusion imaging and pathologic indicators of angiogenesis including microvascular density (MVD) counts and vascular endothelial growth factor (VEGF) expression levels of VX2 carcinoma in rabbit soft tissue. Methods VX2 tumor was inoculated in the muscular tissue of right posterior limbs of 18 healthy New Zealand white rabbits ,on which arterial spin labeling perfusion imaging was performed by using flow sensitive alternating inversion recovery pulse sequence 3 to 4 weeks later. Arterial spin labeling images were transferred to an independent computer for post-processing and blood flow (BF) maps were generated from them. BF values were measured in 2-4 regions of interest (ROIs) for each tumor. The rabbits were sacrificed after MR imaging. The gross specimens of tumors were obtained and tumor tissues were taken from the non-necrotic regions corresponding to ROIs on BF maps. Immunohistochemical staining of the specimens was performed by using CD31 monoclonal antibody to calculate MVD counts, using VEGF antibody to calculate VEGF expression levels. Correlation between BF values and MVD counts as well as between BF values and VEGF expression levels was evaluated using Spearman correlative analysis.Results On BF maps, viable tumor tissues showed high BF values compared with muscle, but there were areas without blood perfusion in some tumors. Under microscope, the microvessels positively stained by CD31 appeared as light brown areas, and the cells positively staied by VEGF showed reddish brown areas within their cytoplasm. Totally 39 pieces of VX2 tumor tissue were analyzeed There was a significant positive correlation between the BF values(M =6. 4 ml · 100 g-1 · min-1 ) and MVD counts(M =6. 8) (r = 0. 906,P ＜ 0. 01 ), and no significant correlation between BF values and VEGF expression levels ( M = 8％ ) ( r =0. 116, P=0.483). Conclusion BF value can be used in evaluating angiogenesis of soft tissue tumor through its reflection of MVD counts, and thus may be helpful in evaluating the prognosis of soft tissue tumor and making plan for their treatment.