With the rapid development of Web 2.0, there has emerged a large number of product reviews written by users with their subjective views on online community, blog and e-commerce website. In product reviews, users are mostly concerned about the comments on a certain aspect or feature of the product, so the fine-grained opinion mining on product aspects is the current research focus. The early researches in this field focused primarily on explicit aspects but neglected implicit aspects. However, the implicit aspects expressed implicitly by a few words or phrases are of equally great importance and also can express the user’s opinions and help us better understand user’s comments. The identification of implicit aspect in product review text is a very challenging but significant subtask for opinion mining. This paper focus on identification of implicit aspects in product reviews. Aiming at deficiencies of previous works, this paper presents a novel association rule set mining approach that considers not only adjectives as indicator for the task, which integrates the improved collocation extraction algorithm with the topic model approach, and extracts reasonable and effective association rules from product review corpus as much as possible through the two complementary approaches, thereby enhancing the effect of identifying implicit aspects. Experiments show that the presented method is effective, and in the case that the precision does not reduce, the overall performance improves by about several percents compared to the baseline approach.