Consumer reviews are dominating consumer purchase decisions, and restaurant (cuisine) reviews are one of the most popular genres of consumer reviews. In contrast to traditional named entity recognition (NER) targets, cuisine names are promising targets that were neglected in previous studies. This study proposes a novel cuisine name extraction algorithm, which can extract cuisine names from restaurant reviews effectively with satisfying performance. Based on the algorithm, we processed more than 12,000 Chinese restaurant reviews and generated a "cuisine map", showing the most popular dishes of restaurants in a map-view to help users make good purchase decisions.