Abstarct: In this dissertation we will study the chaotic time series prediction using phase space reconstruction. The proposed methods for chaotic time series prediction are as follows.We will determine the nearest neighbor point and trajectory after the phase space reconstruction. Then the auxiliary trajectories are determined using linear transformations. The prediction is baxsed on equations called the center of gravity equations. We will use these equations as a model for estimation. The estimation process is performed by the hypothetical movement of particles in neighbor and auxiliary trajectories in accordance with recursive center of gravity equations. The goal is to track the neighbor trajectory so that main trajectory will be predicted simultaneously.