Facing the expanding range of operational risk factors and limited information revealed about listed companies in the market environment competitive fiercely, the auditors need the scientific method to obtain information effectively and evaluate the going-concern problems scientifically. This article will start from all elements which affect the going-concern ability of listed companies, analysis and research the influence on audit opinion of listed companies from these factors, establish prediction model for the types of the listed companies'audit opinion, and attempt to increase the ability of judging the going-concern problems objectively for the CPA. The empirical study about going-concern opinion of listed companies which under a certain condition during the year 2006 to 2008 with the established model in order to regulate the audit conduct for the CPA and provide corresponding theoretical basis for perfecting auditing standards.Based on previous studies, this paper will analyze related theory about going-concern ability and the process for judging the going-concern opinion with the method combining theory and practice. According to the CPA's thinking process about showing audit opinion after deciding the going-concern ability of the audited unites, we found that there exist the correlation between going-concern ability and the types of audit opinion about listed companies. Then we assume that the listed company which have some problems on the going-concern ability will receive the non-standard audit opinion more easily. On this foundation, using the logistic regression analysis, we show the relationship between every index affecting going-concern ability and the types of audit opinion with function. After analyzing empirically, examining the prediction capability about established model, we give an objective and scientific evaluation.The final results confirmed that the important financial indicators of listed companies are related strongly with the types of audit opinion. Second, the non-financial indicators in the model also related with the dependent variable which means the probability of receiving the non-standard audit opinion. Finally, the effect for forecasting the samples of the forecast group is very optimism, the accuracy of forecasting for clean sample is up to 88.23%, for non-clean sample is up to 100%, and for the whole, the accuracy is up to 92.31%. The model we established for forecasting audit opinion not only contain the traditional financial indicators which reflect the going-concern ability of listed companies, but also include two non-financial indicators which explain the going-concern ability of listed companies indirectly, they are if the listed companies occurred asset restructuring or changed the certified public accountants during the year audited. The model added non-financial indicators doesn't weaken the predictive power but increase the forecast accuracy about non-clean sample up to 100% which shows further that non-financial indicators have a important affect on processing of receiving the non-standard audit opinion, and this impact can be forecasted by quantitative method, but not be ignored as non-visualization like financial indicators, and this is the main innovations about this article. |