Leading-edge-groove (LEG) tilting-pad thrust bearings (TPTBs) can be simulated by using Reynolds-model-based thermo-elasto-hydrodynamics (TEHD). The accuracy of predicting the lubrication oil temperature at the leading-edge inlet of the pad has an important influence on the lubrication performance calculation results. The area between the two adjacent pads of the LEG bearing and the LEG together constitute the lubricant thermal mixing zone within the bearing. The oil-temperature prediction model presented in this study is implemented by utilizing artificial neural network (ANN) regression. The detailed steps include determining the design space by employing the design of experiment and establishing a dataset by using computational fluid dynamics. The dataset is subsequently employed to train and test the ANN. The ANN-TEHD model is constructed by inputting the ANN-predicted inlet-oil temperature into the TEHD model. Compared with the results of the TEHD model based on the traditional oil-mixing theory, those of the ANN-TEHD model show better consistency with the experimental data. Thus, this model is proficient in accurately predicting the inlet-oil temperature of LEG TPTBs and facilitating a comprehensive analysis of the lubrication performance of bearings.