Surface net radiation is a key variable of energy and water exchange processes in land-atmosphere interactions. Accurate estimation of spatial-time changes on land-surface net radiation flux is a critical study in global warming for climate change research. This study aims to summarize the three main methods for estimating net radiation at present, to analyze the advantages and disadvantages of the different methods used for surface net radiation, and to propose an effective method using data assimilation to improve net radiation at present and in the future. This paper summarizes the main methods, including observation, simulation, and assimilation, used at present for surface net radiation. Observations include ground-based measurement and remote-sensing estimation. Simulation is mainly concerned with the estimation of land-surface process models. The advantages and disadvantages of observation and simulation are analyzed, and the current status and progress of surface net radiation observation and simulation are reviewed in this study. Moreover, a method of data assimilation is proposed to compensate for the shortcomings of observation and simulation. The advantages and disadvantages of popular assimilation algorithms, such as sequential (Kalman filter algorithm and its derivative, particle filter algorithm) and variational (four dimensional variational assimilation algorithm or 4D-Var) assimilation algorithms, are introduced and analyzed in this study. Furthermore, the research progress and existing problems of assimilation algorithms in surface net radiation flux estimation are summarized. At present, progress has been made in estimating the surface net radiation using data assimilation combined with multisource observations. However, problems exist in these methods. A model is calibrated by the observation data of different sources to obtain the continuous and consistent high-precision surface net radiation prediction value in time and space in the process of data assimilation using different assimilation algorithms. The adaptability of different land-surface process models varies in different research areas based on the simulation results of the model, and the precision and accuracy of the different models in describing the surface water thermal process are significant. Based on the assimilation of remote sensing data, the study of assimilating an optical remote sensing data product has been successful in estimating the net radiation of the surface.