Tropical peat exhibits low shear strength, high compressibility and pronounced strain-softening behaviour, making reliable constitutive modelling essential for the analysis and design of embankments and foundations constructed under short-term undrained conditions. Although the Hardening Soil (HS) and Soft Soil Creep (SSC) constitutive models are widely implemented in PLAXIS, their capability to reproduce the complete undrained stress–strain response of tropical peat has not been systematically evaluated using laboratory-based calibration. This study investigates the performance and limitations of the HS and SSC models through numerical simulation of consolidated undrained (CU) triaxial tests conducted on undisturbed Pontian peat and associated very soft soils obtained from Mukim Rimba Terjun, Johor, Malaysia. Numerical simulations were performed in PLAXIS 2D using laboratory-derived constitutive parameters, followed by iterative calibration and quantitative evaluation using the Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and coefficient of determination (R²). The HS model successfully reproduced the nonlinear pre-peak stress–strain response and peak deviatoric stress mobilisation but was unable to simulate the progressive post-peak strength degradation observed in the laboratory tests because its constitutive formulation does not incorporate destructuration or damage evolution. In contrast, the SSC model provided a smooth viscoplastic response suitable for representing creep-related behaviour but consistently underestimated peak undrained shear strength. The apparent reduction in deviatoric stress predicted by the SSC model resulted from excess pore-water pressure generation and effective stress redistribution rather than true constitutive strain softening. Sensitivity analyses identified the effective friction angle (φ′), reference secant stiffness modulus (E₅₀, ref) and pre-overburden pressure (POP) as the dominant parameters governing the HS response, whereas the compression index (Cc), swelling index (Cs), secondary compression index (Cα) and stress-history parameters controlled the SSC response. The study establishes a laboratory-based constitutive calibration framework that distinguishes parameter-calibration uncertainty from constitutive-model limitations and provides practical guidance for selecting appropriate constitutive models for numerical analyses of tropical peat. The findings demonstrate that the HS model is more suitable for short-term undrained stability analyses, whereas the SSC model is more appropriate for long-term creep and settlement prediction.