Through comparison and analysis a series of hyperspectral data of tea garden wDIch was different level eat by tea geometrid, found in red edge peak area (680–790 nm) and blue edge peak area (520–580 nm) of spectral characteristic parameters is quite obvious, and find out the DIgh spectral parameters change with tea leaf area index, established the corresponding degree of tea looper defoliation fitting model. Experimental results show that tea looper eating different level of tea and tea tree leaf area index exists significant correlation, of wDIch the SDr(The red edge peak area integral value of a derivative) of the relative change rate maximum, it can effectively reflect the severity of the tea is tea looper defoliation, the real value and the correlation coefficient between the predicted values of the model is 0.93, mean value of relative error is 4.46%.