The arbitrary selection of phosphate fertilizer source and their low use efficiency in tobacco production makes it necessary to select rational phosphate fertilizer in purple soil, which is characterized with low soil phosphorus (P) availability. No phosphate fertilizer (control), single superphosphate (SSP), calcium–magnesium phosphate (CMP), ammonium polyphosphate (APP), monoammonium phosphate (MAP), and diammonium phosphate (DAP) were used to study the effects of different phosphate fertilizer source on inorganic P transformation, tobacco growth, and P utilization. During the 90-day incubation experiment, there were obvious differences in the dynamics of soil inorganic P fractions among the five P applied treatments, and the soil Olsen-P contents of all P applied treatments showed an increasing trend while soil pH showed the opposite trend. In the greenhouse pot experiment, the five P applied treatments increased the agronomic traits and P use efficiency (PUE) of tobacco compared to the control, and CMP and DAP had the larger increase of tobacco weight and PUE. Also, the soil Olsen-P, pH, phosphatase activities, and inorganic P fractions obviously differed among the five P applied treatments, the CMP had the highest alkaline phosphatase (ALP) activity and NH₄Ac-extractable P (Ca₈-P) contents, and DAP had the second-highest acid phosphatase (ACP) and neutral phosphatase (NP) activities and highest NaHCO₃-extractable P (Ca₂-P) content. In addition, the differences in the contents and proportions of inorganic P fractions between the indoor incubation soil and the tobacco-growing soil indicated that Ca₈-P was an important source of P for tobacco growth. In all, the CMP and DAP are more suitable phosphate fertilizer source in purple soil, as they better promoted tobacco growth and development and P utilization.