In this paper, we argue that off-balance-sheet information is potentially useful to evaluate accounting information quality in addition to financial indicators. We propose a novel method to evaluate the accounting information quality of listed companies in China by merging four financial attributes and six indicators derived from off-balance-sheet information, where the principal component analysis is used to filter the indicators which count heavily for the evaluation. Validation and comparison are implemented on the dataset of all financial frauds and violations (5031 cases) of 1995 listed companies of Shenzhen Stock Exchange from 2003 to 2017. The results show that the proposed method outperforms both the traditional “DD model” (Dechow and Dichev, The Accounting Review 77(1):35–59, 2002) and the model only involving financial indicators in identifying and predicting financial fraud of enterprises. The results highlight that off-balance-sheet information is indeed useful and important to evaluate accounting information quality.