This thesis applies a new decomposition method proposed by Firpo et al. (2009) in detail and the RUMiC data to estimate the wage discrimination effects to rural migrants in China. We also solve some statistical problems in Wen (2014). We correct the identification problem with Yun (2005's approach and Jann (2005)'s Stata module. And we introduce Fortin (2008's approach to modify the non-invariance problem. According to the empirical results, the trends of discrimination proportion to rural-urban migrants are almost the same in both 2008 and 2009; the average discrimination proportions are 28.7% and 23.2% respectively. Compare with other researches, the main finding of this thesis is the lowest discrimination effects point to rural-urban migrants disappears at the middle or above-middle wage level. Instead, the low and high wage levels groups get more wage discrimination. In general our estimates of the mean discrimination effects are in line with the main trend compare with other studies. We also find the high-paying occupation access barriers for rural-urban migrants, since the occupation taking takes up a large proportion of endowment effects. It is another type of discrimination. Compared the return of education between 2008 and 2009, we could find that rural-urban migrants suffer more discrimination under the financial crisis. The existing researches are almost cross section data with duration of only one year, lacking of time series and panel data. The RUMiC project tries to do this type of follow-up survey and only published the first two waves of China data regarding rural-urban migrants, urban household, and rural household in 2008 and 2009. If we could use more RUMiC data waves to research this topic again, there must be new interesting findings.