Lithium iron phosphate is one of the main industrial raw materials for the lithium-ion polymer battery industry. Compared with traditional raw material lithium cobalt oxides, lithium iron phosphate has significant advantages in specific energy, battery life, cost and environmental compatibility. Lithium iron phosphate, which is carbon composite cathode material for lithium-ion batteries, is selected as the research object. The content of lithium, iron and phosphorus was determined simultaneously by inductively coupled plasma optical emission spectrometry (ICP-OES). Sample digestion, analytical wavelength, parameters of instrument, solvent effect, background interference for analytical wavelength, accuracy and precision of the method had been discussed. Carbon composite lithium iron phosphate contains about 5% carbon. 0.100 g Sample was digested with 5 mL perchloric acid, heated the digestion solution to clarify visual, and continued heated the solution to smoke and repeated the operation twice as the sample could be digested completely. The sample digestion was checked by the three-dimensional (3D) video microscope. The visual clarify solution was checked by 30 magnification times. It showed that there still had an insoluble matter in the solution. After repeated the operation of added water and continue heated to smoke twice, checked digestion solution by 30 magnification times again; it showed that the sample had been digested completely. The digestion solution was transferred to a 100 mL volumetric flask, diluted to volume and mixed. The analytical solution was prepared as followed: diluted the constant volume digestion solution with the rate of 10:100 to a 100 mL volumetric flask. The best parameters of ICP-OES for this method were: High-frequency power, 1.1 kW. Atomizer flow: 0.8 L·min^-1. Analytical wavelength, observe direction & observe height: Li610.365 nm-axial,Fe259.940 nm-radial/12 mm,P178.222 nm-axial,P213.618 nm-radial/12mm. Calibration solutions were prepared according to matrix matching with regent blank prepared with sample simultaneously, the preparation of calibration solution can minimize the solvent effect. For Li 610.365 nm, fast, automated curve-fitting (FACT) was used to eliminate background interference of Ar 610.564 nm. Under the optimal conditions, the analytical ranges were: Li, 1%~9%, Fe, 20%~40%, P, 10%~30%. The coefficient of the calibration curve, r>0.999 0. The relative standard deviation (RSD, n=7) was 0.35 %~1.01 %. The standard recovery rate was 91.2%~112%. Internal round robins (IRR) result was evaluated by F, t-test, the result showed no obvious difference between the test results of two operators. The method has promotional value for application in the lithium-ion polymer battery industry.