Being able to predict bankruptcy can be very valuable for debtors, creditors, shareholders and other stakeholders. Historically, different models that predict corporate bankruptcy have been constructed. Three bankruptcy predicting models are used in this thesis; the models of Altman (1968), Ohlson (1980) and Shumway (2001). The relatively old original models are applied to U.S. listed firms. It became clear that when the original models are applied to a more recent sample of 1999-2013, the predictive power of the models is very low, and bankruptcy is over predicted. In order to be able to use the relatively old models in more recent periods, the results show that the models have to be re-estimated. The original models with the original variables are used, only the coefficients and the interpretation of the outcome of the models change by downside risk. The downside risk models show that especially variables of short term liquidity are more important now a day in predicting bankruptcy than in the original models. After reestimated the models, the accuracy rates of all models increased. Especially applying the downside risk models of Altman (1968) and Ohlson (1980) to the sample of 1999- 2013 result in high percentages of correctly classified observations and high areas under Receiver Operating Curves. The research was carried out with an aim to contribute to the discussion regarding business failure prediction of Pakistani Firms using Altman (1968) Z-Score, Ohlson (1980) O-Score and Shumway (2001) hazard model, its applicability and validity in Pakistani scenario so that to provide an addition into the literature of drivers of the prediction of the business failure in Pakistan. The research was carried on three research questions; can Altman (1968), Ohlson (1980) and Shumway(2001) model correctly forecast the bankruptcy problems in Pakistani environment, if not, can only changes in coefficients of Altman(1968) Ohlson (1980) and Shumway (2001)model’s ratios are sufficient for this purpose, is there is a need of new bankruptcy model for Pakistani firms and if so which of the financial ratios and relationships between financial ratios are useful in forecasting Pakistani business financial failure. This study used data of Pakistani publicly listed companies for the period of January 1999-December 2013 to test the accuracy of Altman’s Z-score, O-score a Hazard model in predicting failure of Pakistani’s companies and for model development. Multiple discriminant analysis (MDA), Logit and Probit are chosen as the suitable statistical technique. Multiple discriminant analysis (MDA), Logit and Probit is a statistical technique that can be employed to classify firms into one or more mutually exclusive categories such as bankrupt and non-bankrupt firms. MDA approach has been the most popular technique for bankruptcy studies as shown in literature. Results show that Altman(1968) model does fairly well for predicting the bankruptcy of a firm, with accuracy ranging from 94.6% to 83.9% but the model tends to misclassify a non-failed firm in to the failed group with error ranging from 61.1% to 91.1%.The re-estimated model does fairly well for predicting the delisting of a firm, with accuracy more than 80% as well as to predict with 100% accuracy for successful firms into the stable group with no error in whole yearly sample which is a high achievement of this model. Proposed model predict failure with overall 80%accuracy. Study findings suggest that results of Altman (1968), OhlsonO-Score and Shumway (2001) hazard model to estimate financial distress of sample firms should be interpreted cautiously in Pakistan, due to its low predication in successful companies. The re-estimated and proposed models give better results.