A novel method to monitoring the internal temperature of a power battery was proposed based on the Electrical Resistance Tomography(ERT) to complement the temperature measurement in online and to insure the safety of the power battery. Through introducing the Wenner-Schlumberger model in the ERT, the relationships between the internal temperatures and the resistivity of power battery were deduced. And by using the smoothness-constrained least-square method to invert the images of multi-electrode tomographic detection data, the internal abnormal temperature regions were monitored dynamically and a test platform was established. To validate the reliability of this new method, an intersection region model was established to analyze the static distribution of internal resistivity of the power battery. Then, dynamic resistivity distribution was analyzed in five different temperatures. The analysis shows that the color ribbon difference in the inversion image increases by 2.5 times in 45.5 ℃ as compared to those of other four tests, especially at coordinates (0.047, 0.000 5), which changes significantly. Finally, dynamic tracking critical temperature in mutational area shows that the resistivity rebounds suddenly at 0.173 Ω·m, which indicates that a mutation occurs in the battery. These results demonstrate that the proposed method not only offers a visual and reliable monitoring result of internal temperature changes for the power battery, but also can provide a new research pathways for the power battery, such as safety characteristics, life prediction, and load control.