【Objective】This paper describes the study of a test plantation of 25-year-old Pinus massoniana (masson pine) trees, in which gene expression and DNA variation were examined on three levels: provenance, location and stand. The purpose was to provide a valuable reference from which genetic breeding strategies can be developed.【Method】The SSR molecular marker was used to perform genetic structure analysis, and the Arlequin 3.5 software was used to estimate the relative weight of each level of variation according to the linear model. The stand growth data were analyzed statistically by the SAS analysis software based on a linear model, and the relative weight of each level of variation was estimated by the four-factor nested model.【Result】Genetic diversity analysis showed that there was no obvious difference between provenances or within the same provenance. The results of molecular ANOVA analysis showed that the genetic variation between provenances accounted for only 2.56% of the total variation, and different locations within a provenance accounted for 1.66%. Alternatively, differences between individuals among those locations tested accounted for 95.78% of the total variation. Analysis of the growth phenotype data also showed no significant difference in diameter at breast height (DBH) and volume character between provenances, between locations within a provenance, or between superior and normal stands within a location. 【Conclusion】ANOVA analyses of molecular and growth phenotype data revealed that the major genetic variation of masson pine in Guangxi exists among individuals within a stand. Therefore, it is better to develop superior individual plant selection criteria in a superior provenance. This conclusion provides valuable information that can be used in the formulation of genetic breeding strategies for masson pine.