For the expenditure on long-term care of the aged is a compound variable composed of several variables, steps of estimating it is proposed in this study. We assumed that the rate of the aged with disability, the growth rate of healthcare professionals' real wage and the growth rate of' CPI could change continuously in the time, and apply GM(1,1) model to forecast the above there variables and the number of the aged. Therewith, we could examine at the same time, which property of the above variables might have an impact on the applicability of GM(1,1). The results illustrate that the growth rate of the total expenditure on long-term care will forecasted to be 208.99% from 2001 to 2011. The main reason for its tremendous increase is the rapid growth of the rate of the aged with disability. In addition, it is found that GM(1,1) performs very well, when being applied to forecasting the sequences with stable and fluent growth, while poor, when being applied to forecasting the sequences with central symmetry.