Based on classes (including large,medium,small) of Tamarix ramosissima crown to select and weight samples from Yingsu to Ka’erdayi Section,the relationship between age and basal diameter of T.ramosissima was analyzed using statistical correlation method.In order to predict the above ground biomass of upright growth and single cluster of T.ramosissima,estimation models were established,then examined the accuracy and filtered the best at last and performed a Chi-Square test between measured value and estimated value of biomass estimation models.The result shows that:(1)After water supply,the basal diameter of sprouting upright growth increased with increasing age,while there was no regularity in the upright growth sprouted before the water supply.In the arid deserts with different groundwater conditions,basal diameter of T.ramosissima can’t be considered as the basis to determine its age.(2)By analyzing and comparing the estimation model which have four kinds of functional forms including linear function,power function,exponential functions,logarithmic functions,the power function model with the following w=2.007(d²h)^0.837 formula was the best,the accuracy reached 89.92%.(3)Single cluster biomass estimation models of T.ramosissima were built taking crown area,crown girth and height as independent variables,respectively,whose estimation accuracy was 78.95% and 80.09%.(4)Taking measured value acquired with harvest method as real value to analyze the accuracy of estimation models,the accuracy of biomass estimation models built taking crown area,crown girth and height as independent variables was 80.54% and 66.85%,respectively.(5)The differences between measured value and estimated value of biomass estimation models were analyzed by Chi-Square test,the result shows that:There was no significant difference between them.Because the crown area of T.ramosissima can be acquired easily form high resolution remote sensing data,the estimation model which take crown area as independent variables can provide a basis for biomass estimation.