Awareness of time occurrence of developmental stages of agronomic crops, such as safflower, is very important. To introduce model for developmental stages of three safflower cultivars, Arak, Sofea and Goldasht by temperature and day length, planting date, trials conducted in Kabootar-Abad Agricultural Research Station during 2003-2009. To estimate duration of each developmental stage, the measured duration of each developmental stage considered as the dependent variable and various temperature variables, day length and combination of them considered as the independent variables in a stepwise regression procedure. A step of regression was considered appropriate if the highest R2 was accompanied by the significant (p £0.05) regression coefficient and partial R2. Days from planting to emergence, emergence to heading, emergence to flowering, emergence to ripening and flowering to ripening were affected by planting dates. By increasing temperature, developmental stage durations decreased. Day length had the most effect on emergence to heading and emergence to flowering duration and by increasing of day length, they decreased. Tmin and T2min were the variables which entered the regression model and explained about 89% of variation of planting to emergence period. Periods of emergence to head visible was explained by DL×Tmax and DL2×T2max in accuracy about %82. DL×Tmax was the only variable which entered the regression model and explained about %91 of variation emergence to flowering period. DL×Tmax and T4mean explained the most variance (84%) of emergence to ripening period. T2min was the only variable which entered the regression model and explained about 60% of variation of flowering to ripening period. Anticipating of heading, flowering and ripening stages are very important for control of fruit fly and providing of water irrigation and harvesting combine respectively.