Abstract: Based on the observed rainfall data, the accuracy of four remote sensing precipitation products, such as TRMM 3B42 V7, CMORPH CRT, PERSIANN and TRMM 3B42 RT, were evaluated in the Lancang River Basin. The evaluation was performed at multiple time scales, ranging from daily to yearly, for the years from 2001 to 2014, but only one spatial scale: grid spatial scale. On the ground of observations and the selected high-accuracy products, the SWAT model was applied in runoff simulation to further explore the applicability of remote sensing precipitation products in the Lancang River Basin. The results show that four products present low accuracy on daily time scale, and their correlation coefficients（R) with observed rainfall data are less than 0.5; as a whole, TRMM 3B42V7 is the most accurate dataset with R being 0.95, 0.90, from monthly to yearly time scale, respectively; the accuracy of CMORPH CRT is slightly lower than that of TRMM 3B42 V7, with R being 0.89, 0.77, from monthly to yearly time scale, respectively; TRMM 3B42 RT and PERSIANN have present poor accuracy, and their R in different time scales are relatively low, with R basically maintaining below 0.5. The runoff simulation results show that the skills of TRMM 3B42 V7 and CMORPH CRT are not as good as that of the observed rainfall data on simulating the daily and monthly runoff of the river basin outlet section, but all three can better reflect the variation trend of measured runoff in runoff simulation with the corresponding Nash-Sutcliffe coefficient（NS) of the three keeping within the range of 0.65-0.94. From the perspective of the NS, the value difference of the three is insignificant. From the perspective of relative error of water quantity（RE), all of them underestimate water quantity , but the RE value of CMORPH CRT（within the range of -10%—-20%) is larger than that of the other two（ within the range of -4%—-10%). Overall, the runoff simulation capability of TRMM 3B42 V7 is better than that of CMORPH CRT in the Lancang River Basin.