Corpus methodology which prevails in metaphor studies in recent years provides empirical support for metaphor research, broadens its scope and to some extent, addresses methodological limitations in previous studies. This research paradigm, however, receives doubts and challenges in terms of data collection and metaphor identification procedures, specifically with respect to data retrieval, and the representativeness, balance, and indirectness of corpus data. This paper systematically categorizes and analyzes these issues and puts forward relevant recommendations for future research.