A fast and accurate temperature field reconstruction method based on the inverse distance weighting (IDW) method and the long short-term memory (LSTM) network is proposed in this paper to solve the challenging task of quickly and accurately obtaining the global temperature distribution during the boiler combustion process. The method combines data from sparse temperature measurement sensors with sensor location information and utilizes the IDW method to interpolate temperature data over the entire boiler wall area. This process helps fill in the areas not covered by the sensor, thereby enabling temperature control and obtaining a preliminary two-dimensional temperature field. Subsequently, an LSTM network is used to further improve the reconstruction accuracy. The LSTM network effectively captures the long-term dependence of the temperature field, accurately predicts temperature range trends and change patterns, and reduces the impact of disturbances such as sensor measurement errors and environmental interference. Experimental results show that this method is superior to mainstream methods in terms of RMSE and MAE evaluation indicators, and shows higher speed and accuracy in reconstructing the two-dimensional temperature field. This method is expected to promote research in the field of boiler temperature and has good application prospects in the field of boiler combustion.