This paper studies the dynamic relationships between Taiwan industrial section's average wages and labor productivity. Box-Jenkins ARIMA and transfer function models are employed in this study. This paper uses monthly data between January 1986 and March 2007 to analyze these time series using classical white-noise regression models. This paper finds that the Taiwanese labor productivity is mainly influenced by the average monthly wages, and also finds that the dynamic relationship between the labor productivity and average wages over time, shifting from a longer lag response to a shorter lag response. This result has important implications for forecasting the average wages and labor productivity. The forecasting value of these models is carefully examined. This paper concludes with comments on possible future research directions in this field.