In this paper, we propose a requirements guided dynamic approach to address software clustering which aims at providing the logically meaningful and high-level decompositions of large and complex systems. In our approach, the hierarchical structure of functional requirements are constructed by a text document clustering technique named Hierarchical Agglomerative Clustering (HAC) as a high-level skeleton to facilitate the further decomposition of source code through dynamic analysis. We also perform an experimental study based on a GNU system and present the quantitative and qualitative analysis of the experimental results.