The objective of this research is to propose a statistical model that could estimate the probability of occurrence of bankruptcy for Internet companies. After the recent and drastic losses of investment capital in companies in this new sector of the economy, financial institutions, individuals and all investors wish to know the real financial position of these companies called dotcom. This research selected American companies that have filed a petition under the United States Bankruptcy Code, between 1999 and 2001, and companies which have not done it, by convenience sampling, that list their shares on stock markets and operate in e-commerce, i.e. companies that sell their products or services through the Internet. The financial statements of these companies were also used to identify,by analyzing a T test of independent samples, the most significant variables for discriminating the two observed groups in the sample: the bankrupt and the nonbankrupt companies. After analyzing the variables statistical distributions, a logistic regression model revealed to be the more appropriate for the research, for not having the multivariate normality assumption. The conclusion of this research proposes a statistical model which indicates the probability of an Internet company becoming bankrupt or not, with a Nagelkerke R Squared of 0,887, and an overall percentage of correct prediction of 97,4 percent. The model uses several variables not previously included in similar previous financial difficulties prediction models.