<p>An innovative experimental scheme is proposed to separate the coupled effects of adiabatic heating and strain rate on Strain-Induced Martensitic Transformation (SIMT) in FA-grade ultra-high-strength steel (UHSS) at high strain rates. This method combines strain-controlled tests (non-isothermal) and strain increment tests (quasi-isothermal). Conventional split Hopkinson tensile bar (SHTB) devices face difficulties in performing strain increment/controlled tests at high strain rates due to interference from multiple impulse loadings. A novel SHTB device, incorporating a dual-momentum capture bar and sleeve structure, enables reliable single-pulse dynamic loading in both test modes. Quantitative analysis was conducted to compare the relative influences of adiabatic heating and strain rate on martensite suppression, supplemented by microscopic characterization. The results indicate that the martensitic phase transformation volume fraction increases with strain but decreases with increasing strain rate in both test modes. Adiabatic heating was identified as the dominant factor suppressing SIMT (accounting for 44.6% of the inhibition at a strain of 0.18), while strain rate played a secondary role. Specifically, the attenuation of adiabatic heating delays fracture occurrence. Microstructural analysis revealed that the difference in martensite transformation content between the two experimental modes arises from varying degrees of dislocation accumulation at phase interfaces. Adiabatic heating causes the phase transformation path to deviate from the conventional orientation relationship (OR) matching pattern. The effects of strain rate and adiabatic heating on interphase strain partitioning were found to be negligible, establishing SIMT as the primary regulator of macroscopic mechanical responses, with interphase coordination playing a secondary role.</p>