The four components of temperature (max., min.) phenomena, seasonal(S), Trend(T), cyclical(C), and random (I) for Tartous city have been studied. Four different methods (Average percentages method, Percentage of the general trend method, The ratio of the moving average method, Link Relative method) are used to deduct the (S) components and seasonal index for each method is determined. The statistical inferences pointed that the Average percentages method can be used in the prediction of temperature. for the year 2003 depending on a historical record (1957- 2002). The result of this deduction showed that the temperature is a cyclical phenomena. The known statistical test like mean, Standard deviation and cumulative probability have been done which showed a good correlation between the predicted and historical data.