The absorbance of lead in flour was determined by microwave digestion method and atomic absorption spectrophotometer. Box-behnken Design was used to determine the best instrumental analysis conditions. Through the single-factor test, the four factors of ashing temperature, ashing time, atomization temperature and atomization time were selected for the three-level test. The mathematical model was established by Box-Behnken Design, and the optimal conditions were sought by variance analysis, response surface analysis and desire function optimization. Results show that the established mathematical model of the R²=0.9671，R²_Adj =0.9342 shows that the model has good credibility, this model could be used to predict the coefficient of variation CV % = 3.08%, indicates that under the condition of the experiment, variation is small, the model has good repeatability, when ashing temperature is 525℃, ashing time is 20s, atomization temperature is 1550℃, atomization time for 5s absorbance maximum 0.24; Therefore, this study can provide a good theoretical basis for the determination of lead by atomic absorption spectrophotometer.