I have got a Problem, I am developing an Application which should be able to do some mathematic calculations. These calculations have to be exact (or rather not obviously wrong)
This is a consequence of the IEEE 754 floating point representation, not an error. To deal with it, round your result to an appropriate precision.