In order to identify the relationship between the major physical and chemical indexes with the smoking quality of the primary cured tobacco leaves,the sensory quality was predict.A total of 222 tobacco samples were collected and tested in Wenshan,and the typical correlation and multiple regression analysis were carried out.The results showed that the combination of mellowness,aftertaste,flavor quality and volatility had a strong correlation with leaf color,oil and identity,and the cumulative contribution rate of the first two pairs of canonical variables reached 98.13%.The comprehensive indexes of mellowness,irritation,aftertaste,volatility,offensive taste,flavor quality and dryness are correlated with stem content,leaf width and equilibrium moisture content.The cumulative contribution rate of the first two to the typical variables is 99.04%.The combination of flavor quality and volatility was also correlated with potassium,ratio of reducing sugar to total sugar and nicotine in chemical composition.The cumulative contribution rate of the two pairs of effective typical variables was 93.29%.The equation established by stepwise regression analysis showed that the descriptive of smoking quality showed positive correlation with maturity,equilibrium moisture content,stem content,total plant alkali,reducing sugar,ratio of reducing sugar to nicotine,ratio of total nitrogen to nicotine,ratio of reducing sugar to total sugar,and K/Cl ratio.In conclusion,according to the regression equation between physical and chemical indexes and sensory assessment quality obtained by canonical correlation analysis and regression analysis,sensory assessment quality can be comprehensively predicted.