Dear Hazar,

if you don't remove nonrandom components like trend or seasonal from your time series, you are going into a great risk to found out something unusual and unexpected because of possible effects of spurious correlation (in the case of existing trends), or overstating or understating the tightness of the correlation (in seasonal components case). So you'd better to eliminate trend and seasonals first.

You need to run co-integration and causality tests for stationary data

Hello you could use ARIMA models for time series analysis. Try eviews statistical package. If you forecast several time series variables, you could use ithen as independent in a regression model.

Thank you very much for your enlightenment.