The performance of lithium-ion batteries (LIBs) is sensitive to the operating temperature, and the design and operation of battery thermal management systems reply on accurate information of LIBs' temperature. This study proposes a data-driven model based on neural network (NN) for estimating the temperature profile of a LIB module. Only one temperature measurement is needed for the battery module, which can assure a low cost. The method has been tested for battery modules consisting of prismatic and cylindrical batteries. In general, a good accuracy can be observed that the root mean square error (RMSE) of esitmated temperatures is less than 0.8 °C regardless of the different operating conditions, ambient temperatures, and heat dissipation conditions.