In this thesis, we provide a review of the theory of reproducing kernels in Hilbert spaces, and a brief survey of its applications in the study of analytic function spaces and in time series analysis. The principal aim of this thesis is to provide a theoretical basis for time series analysis that is independent of the commonly employed stationary hypothesis.