Leaf area index (LAI) is one of the most important structural parameters of terrestrial ecosystems. In recent years, the use of LAI-2200 to measure leaf area index of grassland has gradually increased, but few studies have evaluated the accuracy of the optical method to estimate LAI in grassland ecosystems. In order to validate the measurement accuracy of LAI-2200 for grassland leaf area index, the effective plant area index (PAIe) retrievals by LAI-2200 were compared with the simulated LAI-2200 measurements and the plant area index (PAI) retrievals by destructive sampling in 19 grassland sites. As an additional biophysical parameter of comparable importance to LAI, the clumping index (CI) of grassland was also analyzed. The PAIe retrievals from LAI-2200 4ring data with smaller zenith angle perform better than from LAI-2200 5 ring data, and correspond very well with the destructive PAI values (R2=0.951). Reasonable agreement of the PAIe retrievals from LAI-2200 with the destructive sampling and the simulated results verifies the reliability of LAI-2200 used in sparse grassland. PAIe is slightly less than PAI when PAI is less than 3. As PAI continues to increase, PAIe has been seriously 〖JP2〗underestimated due to clumping effect: CI=0.97 for PAI < 3, CI= 0.88 for 3 6. Therefore, the LAI retrieved from LAI-2200 with the assumption of the random foliage distribution might yield inaccurate results in clumped grass, which need to be corrected using CI values.