Growth in agriculture production can be attributed to increase in inputs, institutional reform and the growth of total factor productivity. However, the development of the agricultural economy in China is believed to increasingly rely on the total factor productivity (TFP) growth because the increase in inputs is has limited potential to continue infinitely (Rozelle et al., 2010). In this thesis, the major purpose is to have a better understanding of the total factor productivity growth in China's agriculture in the 2001-2012 period, and compare with that in the late 90s. To reach this objective, we need to take three steps. First, we collect the data of main inputs and output for 21 China's agriculture commodities from the National Cost of Production Data Set, which is thought to contain the most complete data of farm input and output in China (Rozelle et al., 2010). Second, based on the fact that compared to data envelopment approach (DEA), stochastic frontier approach (SFA) has the advantage that it attempts to account for data noise (data errors, omitted variables, etc.) and allows us to use standard statistical test to test the significance of variables included in the model (Hossain et al., 2012), we choose to use SFA to estimate the TFP growth rate for each commodity on provincial and time basis. Finally, we decompose the TFP changes into two main components: technical efficiency changes and technical progress. To better analyze the TFP growth and its decomposition, the commodities are divided into three categories according to their nature. The estimation indicates that the growth rate in China's agricultural TF for all 21 commodities has dropped to 1.4% in the study period, which is below 2% per year, a rate that is thought to indicate if an agricultural sector is healthy or not (Rozelle et al., 2010). Growth of TFP for the grain crop commodities grow at a negative rate of -0.33%; TFP growth for most vegetable and fruit commodities behave better, whose growth rate is round 3.03%. TFP growth for the last category is 1.52% annually. Another finding is that compared to the result in the 1995-04 period, although technical change accounted for most of the change, the contribution of technical efficiency change had revised to positive, and the technical progress in the study period has decreased over 3% from that in the 1995-04 period. Combined with the decrease of rural labor inputs in the agriculture sector across year, the result implies that we should pay more attention to the effect of factor inputs to the output growth