In order to evaluate comprehensively lettuce leaves phosphorus content and precisely control phosphorus fertilizer quantity, the specific aim of this study was to attempt a strategy for measurement of phosphorus content integrating spectroscopy together with synergy interval partial least square (siPLS) and back propagation artificial neural network (BPANN). Leaves reflectance was acquired with a Fieldspec 3 spectroradiometer that provides measurements in the 350~2500nm spectral range, and then five points smoothed and first order derivative transform were used to eliminate noise effects. siPLS was used to search for the optimal spectral intervals, which corresponded to 950~1070nm, 1430~1549nm,1906~2025nm and 2144~2263nm. 63 wavelengths were selected from 350~2500nm and 25 wavelengths were selected from four optimal intervals by successive projections algorithm (SPA). Principal component analysis (PCA) was implemented on the spectra intervals or variables, finally 7, 4 and 5 PCs were obtained. The siPLS+BPANN, SPA+BPANN, siPLS+SPA+BPANN models were achieved when the number of neurons in the hidden layer was 7, 5 and 3. It was conclude that spectroscopy combined with siPLS+SPA+BPANN were feasible to measure phosphorus content in lettuce, which had better performance than others model, correlation coefficient for the prediction set was 0.911, root mean square error of the prediction set was 479mg/kg.