The failure rate of small and medium enterprises (SMEs) in Sweden is high, with about 6,000 SMEs claiming bankruptcy every year. This paper attempts to identify the main prediction variables that are believed to forecast the failure of Swedish SMEs. The research is principally based on an analysis of a panel data sample consisting of 1,412 bankrupt and 3,084 non-bankrupt Swedish SMEs for the period 2004 to 2006. The statistical technique of logistic regression model is employed to analyse the data. The results, which have a high rate of accuracy, indicate that a set of six variables are significant as bankruptcy predictors: the ratio of short-term debt to total assets, total leverage (the ratio of short- and long-term debt to total assets), change in total assets from the previous year, firm size (natural logarithm of sales), financial expenses to total debt, and return on assets.