This paper employs Bayesian dynamic linear forecasting techniques to investigate the factors driving the predictability of Australian stock market. The unanticipated components of a set of economic and financial variables are chosen as the proxies for the economic risk factors that influence the industrial stock returns. The prior information is incorporated with the predictor variables and updated at each month during the sample period. The final test result reveals that the unanticipated components of term structure and short-term interest rate are the most significant variables to be priced in industry returns. The aggregate dividend-yield variable has influence on some of the industries. The industrial return's predictability is well explained by the time-varying risk premium of economic factors. The comparison between multivariate analysis and univariate analysis strongly indicates that the correlations within the industries are critical in the investigation of the predictability of returns.