Postharvest maturity index for mangosteen (Garcinia mangostana L) is very important for picking and grading during postharvest processing. Skin color change is the primary maturity index for mangosteen. However, determination using human eyes needs many skilled labours and is inconsistent. Therefore, new method in image processing technology using SVM (Support Vector Machine) was employed in this study. Chemical analysis of mangosteen was performed and used as a reference of SVM method. The chemical analysis of mangosteen showed that anthocyanin content increased from 126.20 ppm at stage 1 to 213.98 ppm at stage 6. Reducing sugar content increased from 3.17% at stage 1 to 7.92% at stage 6. The same pattern was found for total soluble solid, an increase from 3.86% at stage 1 to 7.81% at stage 6. Whereas for total acid content and hardness the pattern was the opposite. Total acid content was decreased from 1.78% at stage 1 to 1.06% at stage 6 and the fruit hardness of mangosteen was also declined, showing the number from 4.30 N at stage 1 to 0.69 N at stage 6. For SVM method, image aquisition was conducted for mangosteen images from stage 1 to stage 6, followed by color feature extraction for each stages. The result was trained and tested using SVM and resulted accuracy level of 83.3%.