Objective:To evaluate the influence of reagent blank on the detection limit of the method and the accuracy of results during the determination of lead in foods by graphite furnace atomic absorption spectrometry(GFAAS).Methods: Through the analysis of the reagent blank and spike recoveries prepared by dry ashing,the influence on the detection limit and accuracy of results were evaluated.Results: When reagent blank was used to estimate the detection limit,the detection limit of method increased from 0.004 mg/kg to 0.10 mg/kg.The recoveries of high level concentration ranged from 90.4% to 103.5%,and that of low level concentration ranged from 73.2% to 122.6%.The accuracy of single low content testing result was worse.Conclusion: The detection limit could be reduced because of the existence of reagent blank,and the testing results of low content were strongly affected.Determination quality could be improved through recognizing reagent blank and establishing the blank control chart objectively.