The prime objective of this thesis is to give an overview of the main methods used to decipher polyalphabetic ciphers, using statistical language properties. At the beginning, the basic concepts encountered are defined and the main methods of encryption of classical cryptography are described. Before analyzing polyalphabetic systems, a frequency analysis of a general text was carried out. The results of the analysis give us information which we can use in every encryption system. They are useful in determining the language used, the general system, in reconstruction of the key and the reconstruction of the plain text. Standard or normal distribution of frequencies of letters was also defined. The impact of text volume on the statistical analysis was shown. While analysing polyalphabetic systems with repeating keys and standard cipher alphabets two methods were described. The Kasiski method, which uses factorization and Friedman’s index of coincidence. Both methods use cyclic phenomena which manifests in these systems. Furthermore, in systems with mixed cipher alphabets, a method is described which mostly relies on frequency analysis of digraphs, trigraphs and others, and on the principal of direct symmetry of positions. The applications of the methods are illustrated using examples which were mostly of a military nature. At the end, aperiodic systems and methods with which we can extend encryption keys are described. Such systems avoid the occurrences of cyclical phenomena that facilitated the solutions of previous systems. Using longer keys has been shown to provide a more secure cipher. The probable word method, with which we solved such a cryptogram, is also described. For the one-time pad system, we proved that it is perfectly secure system, if it respects certain conditions.