Asymptotical properties of testing error probability estimates for information security systems are studied under conditions of mutual independence of tests. The information security system’s state is described by three parameters. The first parameter is a real workability, the second one is a result of the workability test, and the third parameter is the number of an information security tool being tested. Mean value and variance of the error probability estimate are determined on the base of the polynomial distribution of error estimates for each test. It is proven that the testing error probability estimate asymptotically has a normal distribution with definite parameters of mean value and variance. Knowing the law of the testing probability estimate distribution and its parameters, we can find a confidence interval for the needed testing error probability for information security systems.