Financial crisis and its consequences are visible in the capital adequacy of many commercial banks, which indicates that the approach banks took to assess credit risk was not sufficiently sophisticated. This article discusses practical methods of insolvency risk modeling for enterprises. In this paper authors analyzed the accuracy of ten models developed by foreign authors to assess insolvency risk, which were validated on the database of Latvian companies. Authors have shown that models developed on historical data for foreign companies are less accurate than the model developed on the basis of financial indicators of Latvian companies. Authors developed a three-factor model that estimates probability of default of Latvian enterprises based on historical data for 2800 enterprises using Binary Logistic Regression analysis. This model can be used to assess the solvency risk as well as PD of enterprises in the calculation of credit risk capital requirement according to Basel II methodology. The model is also applicable in the loan allocation process.