The new Basel II Capital Accord incorporates an explicit capital requirement for operational risk into its proposed capital framework. Unfortunately, although the advanced approaches for the measurement of operational risk evolve rapidly, the absence of reliable internal operational loss databases in many financial institutions is likely to hinder the use of these models. As there is a much greater variety in credit risk modeling approaches, this paper explores the possibility of applying a properly modified version of CreditRisk+, one of the most popular credit risk models, to operational loss data. To assess its usefulness, our this paper investigates the fitting quality of an adapted version of CreditRisk+ on simulated databases generated from three known distributions with thin, medium and fat tails. Our results show that our adapted model, OpRisk+, is able to work out very satisfying Values-at-Risk at 95% level as compared with a lower bound issued from the IMA approach of Alexander (2003), and with the sophisticated approach advocated by Chapelle, Crama, Huebner and Peters (2005). The OpRisk+ approach proves to be extremely worthy in the case of small samples, where more complex methods cannot be applied. This suggests that OpRisk+ can be used in many instances to fit the body of the distribution of operational losses up to the 95% percentile, and Extreme Value Theory or external databases should be used beyond this quantile.