Phosphorus (P) is recognized as one of the most limiting nutrients for primary production in tropical forests because much of the soil P stocks are geochemically bound to iron and aluminum oxides in forms that are largely unavailable for plant uptake. The availability of P to microorganisms and plants can be assessed by evaluating different soil P fractions, and soil P occurs in both organic and inorganic forms, which differ in behavior, mobility, and bioavailability (i.e., labile P, moderately labile P, non-labile P). Labile and non-labile P fractions can serve as sources or sinks of available P, and non-labile P can be released into soil solutions as available P through desorption when the content of available P decreases in the soil. Therefore, it is critical to understand the transformation between different soil P fractions when developing best practices for fertilizer management that aim to enhance the P use efficiency of agricultural systems. Soil microorganisms are important mediators of P mineralization and immobilization, owing to their ability to reduce P availability by immobilizing P in their biomass, which subsequently increases stocks of microbial P in the soil. In addition, lower available P content was observed in tropical soil, accompanied by higher soil phosphatase activity. Furthermore, arbuscular mycorrhizal fungi (AMF) can take up P from pools that are normally considered unavailable to plants. However, the relationship between the dynamics of soil P forms, soil properties, and soil microbial communities has yet to be established in rubber plantation at different ages. Therefore, we compared the soil P fractions (i.e., labile P, moderately labile P, sonicate P, Ca-P, and occluded P) and soil microbial community composition of 4-, 15-, and 31-year-old rubber plantations on Hainan Island, Southern China. The aims of this study were (1) to determine whether the soil P stock and fractions would change with increasing plantation age; (2) to test whether the composition of soil microbial communities would change with plantation age; and (3) to quantify the relationships between P fraction, microbial community, and other soil attributes. The results could be used to develop recommended P fertilizer management practices for sustainable rubber plantations. We found that occluded P was the largest phosphorus fraction, followed by moderately labile P, labile P, sonicate P, and Ca-P in all plantations. The levels of bacteria and both gram-positive and gram-negative bacterial Phospholipid fatty acids (PLFAs) were highest in the 15-year-old plantation, whereas the levels of actinomycetes, arbuscular mycorrhizal fungi (16:1ω5c), and fungal PLFAs were lowest in the 31-year-old plantation. The level of soil acid phosphatase activity did not change with increasing plantation age, and redundancy analysis indicated that both biotic and abiotic factors were important contributors to variation in soil P fractions. Accordingly, the present study demonstrates that the age of rubber plantations significantly affects the soil physicochemical properties and microbial communities of rubber plantations and, consequently, soil P availability and cycling.