Consumer preference is a hot topic in the domain of marking management and e-commerce. Many previous studies have been conducted in this field. Whereas, there are rarely studies building on the particular commodity such as laptop. Therefore, this study explores comprehensive features that affect consumer preference for laptops by mining the online reviews. Firstly, we collect 6531 online reviews for Acer laptop from Amazon.cn and code these reviews with Nvivo10. Secondly, we develop a feature-based consumer preference model named MCPL based on the review text analysis. Considering the data imbalance of the collected 6531 product reviews, we adopt a random cluster sampling method to extract 50 groups with 100 samples per group. Then the correspondent regression analyses are conducted for the 50 groups of reviews. Finally, the meta-analysis is creatively conducted to integrate the multiple liner regression results of different groups. According to the result of meta-analysis, we demonstrate dominant features on behalf of the consumer preference of laptop and draw practical implications for enterprise competition strategies to facilitate product design or improvement.