Abstract: Two provenance trials of Casuarina equisetifolia were conducted at Dongshan of Fujian Province and Yangxi of Guangdong Province in south China.The two trials were 49 seedlots with seeds from Australia Tree Seed Center.Dongshan trial was 40 seedlots which include one local seedlots,and Yangxi trial was 42 seedlots.A complete randomized block design was employed for each site,with 24 trees per plot and 4 blocks or replicates.The 33 seedlots of the same kind were used in the two trials.From 1 to 4 years old,tree height (H) and diameter breast high(DBH) were measured,and single tree volume(V) and survival rate of tree (S) were calculated.Some qualitative traits on tree stem and branch feature were observed which include stem form axis persistence (SFA),stem form straightness (SFS),permanent branches density (PBD),permanent branches thickness (PBT),permanent branches angle (PBA) and permanent branches length (PBL). The results showed that there were significant difference (P0 01) in quantitative and qualitative traits between sites,between provenances,between provenance×site (G×E) interaction,which indicated that those provenances were not only genetic variation in tree growth indexes at each site but also genotype×site interaction.Provenance broad heritability (H 2) were calculated.Using H,DBH,V,S,SFA and SFS,by the multiple objective strategic decision analysis method,provenances in each trial was optimized and ranked.Ten seedlots which was 25% of the total seedlots at Dongshan,and 8 seedlots which was 20.93% of the total seedlots at Yangxi,were screened out as better Provenances.These seedlots performance was better than the local control seedlot.The genetic variation coefficients an broad heritability of provenance in all quantitative and qualitative traits were estimated from 1 to 4 years after planting.Possible suitable areas of C.equisetifolia provenances were predicted in southern China by means of Booth's model.