Pseudorandom number generators (PRNG) are so important in cryptography .Their main advantages are speed, efficiency and reproducibility, so the same sequences of keys (gamma) could be generated at both sides i.e. the sender and receiver to be used during encryption process. But their main disadvantages are periodicities and predictability of random numbers (keys) based on the knowledge of preceding sequences, which result in low level of security when they are used in cryptography. This research shows a method of providing PRNG without the properties of periodicity and predictability, i.e. secured cryptographic pseudorandom number generator, by the usage of deterministic genetic algorithm. This article studies the properties of uniformity, randomness and independence, between two sequences of random numbers; the first sequence is generated by using traditional pseudorandom number generator (PRNG), while the second one is generated by the help of cryptographic pseudorandom number generator which is modified by genetic algorithm (GA). This work shows the graphical and statistical tests which include: frequency test, runs test, Autocorrelation test and entropy. The tests are performed and implemented by the help of three programs: MATLAB, Minitab and IBM SPSS Statistics. The conclusion of this research shows how genetic algorithm would provide cryptographic pseudorandom number generator by improving the quality level of randomness of random numbers generated by conventional pseudorandom number generator (PRNG).