Abstract: Pressure sensors, as one of the core components in modern sensing technology, hold significant application value in industrial automation, medical monitoring, and intelligent equipment. However, their output signals are susceptible to environmental temperature drift, which severely limits measurement accuracy in industrial scenarios. This study proposes a novel temperature compensation algorithm: a Random Forest-based temperature compensation method. By introducing interaction terms between temperature and sensor output voltage as feature inputs for the Random Forest, the algorithm enables more precise decoupling of temperature effects on pressure measurement results. Simultaneously, grid search is employed for hyperparameter optimization to minimize model prediction errors. Experimental results demonstrate that the maximum full-scale relative error of the pressure sensor calibrated by the Random Forest algorithm is 0.13% FS, representing improvements of 48% and 70.5% compared to the Sparrow Search Algorithm-optimized Support Vector Regression (SSA-SVR) model (0.2501% FS) and Particle Swarm Optimization-based BP Neural Network (PSO-BP) model (0.44% FS), respectively.