In this experiment, different multivariate statistical analysis methods were used to determine the effective traits on artemisia yield and essential oils content. After collecting artemisia from Ghom province, Iran, the morphology of plant was studied on 20-40 plants in each plot. Proline, chlorophyll, soluble sugars, sodium, potassium, calcium and magnesium contents were also measured. Results of Pearson correlation indicated that flowering shoot yield had positive significant correlation with total shoot yield (r=0.0.90**), total dry weight (r=0.89**), total chlorophyll (r=0.67*), chlorophyll a (r=0.69*) and potassium (0.89**), and negative significant correlation with soluble sugars (r=-0.68*) and proline (r=-0.81**). Essential oil percentage was significantly correlated to Na content (r=0.74*). Results of principal component analysis showed that the four first factors contributed to 52.24%, 29.93%, 8.75% and 4.08% of the variations. Stepwise analysis also indicated that Na, Ca, Mg and soluble sugars entered to the model respectively. In path analysis, essential oil yield was considered as the dependent variable. Na content had direct positive effect (4.483) on the essential oil percentage. This trait indirectly affected essential oil percentage through Ca (-1.653), Mg (-1.928) and soluble sugars (-0.157).
