Forest net primary productivity (NPP) shows the capacity of forest vegetation to convert or fix CO_2 into compounds. As a reflection of forest carbon sequestration capacity, NPP is one of the main indicators of forest succession and carrying capacity of terrestrial ecosystems. Methods based on remote sensing or inventory data are limited by observation time, which makes the estimation of NPP over a long time period very difficult. However, tree ring data effectively reflect the history of forest growth year by year at decadal and centennial scales, and thus have great advantages in retrieving NPP of forest populations or communities in the past. It can be used to carry out long-term, dynamic, and continuous estimation and prediction of forest biomass, NPP, and dynamic status of carbon cycle. This article summarizes the main scientific research for reconstruction of regional NPP based on tree ring data. The results show that current studies mainly focus on two types of methods for reconstructing the NPP of forests: the first is to obtain forest growth volume in historical periods directly by measuring the radial growth and diameter at breast height of trees; the second is using correlation between vegetation index and tree ring index to infer the evolution of regional NPP in the past. The first approach can be further divided into two applications. Combined with survey data from sampling plots, NPP of the ecosystem can be evaluated by calculating total biomass and productivity of all single trees on the plot. In addition, models that contain tree ring data as an argument that can refer to the vegetation growth can be used to derive NPP of the ecosystems. Measuring forest annual biomass with tree ring data at plot level generates high precision results and the method is suitable for reflecting NPP change of small-area forest stands. For large-scale applications, however, as a result of widely spaced sampling sites this method produces low precision estimates of NPP. The second method is capable of dealing with high heterogeneity in NPP and vegetation indices, and has irreplaceable advantage in estimating forest biomass and NPP at large scales. In some areas, however, the method has limitations because of low correlation between tree ring index and vegetation indices. Accuracy of the results is also limited by the accuracy of remote sensing vegetation indices. Most of the results show that climate factors have conspicuous correlation with NPP in most places. Since the industrial revolution, the warming of the global climate and increasing CO_2 concentration led to the increase of NPP of forest in most areas, especially in high latitudes in the northern hemisphere. These results are beneficial to ecosystem carbon sink and carbon balance evaluation in the climate change scenarios. However, assumptions and limiting factors are involved in the two methods. The spatiotemporal accuracy of NPP estimation using tree rings can still be improved in future research.