The aim of this study was to investigate the effect of rootstock and shoot types on in vitro mangosteen micrografting. The experiment was arranged in completely randomized design (CRD) with two factors. The first factor was the rootstock type, i.e. rooted planlet from the germination of quartered seed, and rooted planlet from the germination of undivided seeds. The second factor was the developmental phase of scion, i.e. dormant buds, and flush (had new leaf more than 2-4 mm). The results showed that rootstock derived from the germination of undivided seed had a higher success rate than other treatments on all variables, except for number of new leaves. The use of flush as scion was better than dormant buds; flush resulted in a higher percentage of successful micrograft and longer shoots. In vitro micrografting had a better growth rate than grafting at the same age. The results of anatomical observation conducted at four months after micrografting demonstrated that there was a good graft union, indicated by excellent fusion between rootstock and scion xylem tissues.