it is desirable for all related insiders and outsiders to discern all potential risk in advance . this paper, with the adoption of special treatment resulted from abnormal financial position as the indicator of financial distress, the univariate variable analysis and multiple variable analysis as the research approach and some financial ratios as variable, tries to find an optimal financial distress prediction model of chinese manufacturing listed companies based on public accounting data . our finding demonstrate that five general financial ratios and three ratios concerning the cash flow have better predicting ability, the erroneous classification ratio are low . these five general financial ratios are earning per share, return on net assets, return on gross assets, growth rate of net profits, growth rate of net assets; the three ratios concerning the cash flow are net cash flows from operating activities per share, net re-earnable cash flows / current liability, net cash flows from operating activities / net profit 研究結(jié)果表明,在單變量分析中,每股收益、凈資產(chǎn)收益率、總資產(chǎn)報(bào)酬率、凈利潤(rùn)增長(zhǎng)率、凈資產(chǎn)增長(zhǎng)率這5個(gè)財(cái)務(wù)比率的錯(cuò)分率較低、預(yù)測(cè)能力較強(qiáng);經(jīng)營(yíng)活動(dòng)凈現(xiàn)金流量與凈利潤(rùn)之比、每股經(jīng)營(yíng)現(xiàn)金流量、可重復(fù)賺取的現(xiàn)金凈流量與流動(dòng)負(fù)債之比這三個(gè)現(xiàn)金流量財(cái)務(wù)比率對(duì)于預(yù)測(cè)上市公司財(cái)務(wù)困境具有有效性;多變量分析中,應(yīng)用費(fèi)雪判別分析和典則判別分析得到兩個(gè)判別模型,在典則判別分析中,應(yīng)用兩種方法確定所建模型的最佳分界點(diǎn),檢測(cè)證明應(yīng)用所得兩個(gè)判別模型進(jìn)行財(cái)務(wù)困境預(yù)測(cè)的準(zhǔn)確率很高。