This thesis introduces a series-wound framework to generate a large number of new one dimensional (1D) chaotic maps based on a combination of two 1D chaotic maps (called seed maps). Examples and experimental analysis demonstrate that the newly generated chaotic maps have more parameters, larger chaotic ranges, and better chaotic behaviors than their corresponding seed maps. And by using the new chaotic map that generated by the serieswound framework, a new image encryption algorithm based on substitution-permutation network (SPN) is also introduced. Simulation results and security analysis to the results show that the proposed algorithm can encrypt different types of plaintext images into random-like, uniform-distributed ciphtertext images that can resist many kinds of attacks.