Magnetorheological (MR) dampers are widely used in the engineering field due to their excellent damping adjustability. However, the performance degradation due to temperature sensitivity must be addressed as a critical issue. This paper presents a thermodynamic modelling approach for double-tube magnetorheological dampers based on stochastic uncertainty theory. Firstly, through mechanical property tests, the damping force-displacement curves at different temperatures were obtained. Experimental results demonstrate that when the outer wall temperature increased from 27.9 °C to 100.92 °C, the maximum damping force decreased by 32.96%.Then, based on the structural characteristics, a heat transfer equation was established. In view of the uncertainties of the physical property parameters of the magnetorheological fluid and the key geometric parameters, the random factor method was adopted to construct a random thermodynamic model. Finally, the predictions were compared with the test results. The improved model exhibited an 18% enhancement in goodness-of-fit compared to conventional thermodynamic models, thereby enabling more accurate characterization of the temperature rise phenomenon. This model serves as a theoretical foundation for thermal stability assessment and temperature compensation control of magnetorheological dampers under complex operating conditions.