Surface albedo significantly affects the balance of surface energy and variation of global climate. Previously

surface a lbedo products have achieved significant success; however

many problems remain

such as low temporal resolution and missing data. Based on the observation data of uniform sites and MCD43B3 production of MODIS

this article verifies three algorithms and their products that arrive separately from a surface albedo production called Global LAnd Surface Satellite（ GLASS）: AB1（ABD01）

AB2（ABD03）

and STF（ABD06）. Then

we analyze and compare the improvement in quality between the initial and f inal products of GLASS. The result shows that the product ABD06 has the best matching with ground observation data

where the RMSE is 0. 050 and R2is 0. 9. In addition

compared with MCD43B3’s surface albedo product

the best consistency is also observed for ABD06

which is supported by the largest number of valid data among the three different products（ ABD06 is 5507

ABD01 is 4290

and ABD03 is 3296）

and the RMSE is better than 0. 03

while R2is 0. 9. Therefore

ABD06 is suitable for the production of long time series’ surface albedo products. The statistical result for a large amount of data also shows that with up to80% of the d ata coming from GLASS’s three products

we can satisfy the accuracy requirement for the deviation between the climate model of the surface albedo and products（0. 02—0. 05）.