The development of InGaAs/InP single-photon avalanche photodiodes （SPADs） necessitates the utilization of a two-element diffusion technique to achieve accurate manipulation of the multiplication width and the distribution of its electric field. Regarding the issue of accurately predicting the depth of diffusion in InGaAs/InP SPAD， simulation analysis and device development were carried out， focusing on the dual diffusion behavior of zinc atoms. A formula of to quantitatively predict the diffusion depth is obtained by fitting the simulated twice-diffusion depths based on a two-dimensional （2D） model. The 2D impurity morphologies and the one-dimensional impurity profiles for the dual-diffused region are characterized by using scanning electron microscopy and secondary ion mass spectrometry as a function of the diffusion depth， respectively. InGaAs/InP SPAD devices with different dual-diffusion conditions are also fabricated， which show breakdown behaviors are well consistent with the simulated results under the same junction geometries. The dark count rate （DCR） of the device decreased as the multiplication width increased， as indicated by the results. DCRs of 210⁶， 110⁵， 410⁴， and 210⁴ were achieved at temperatures of 300 K， 273 K， 263 K， and 253 K， respectively， with a bias voltage of 3 V， when the multiplication width was 1.5 μm. These results demonstrate an effective prediction route for accurately controlling the dual-diffused zinc junction geometry in InP-based planar device processing.