A molecular dynamic computer experiment was performed on a system of 108 particles composed of a single polymer chain and solvent molecules. The state considered was in the immediate neighborhood of the triple point of the system. The polymer itself is an analog of a freely jointed chain. The Lennard-Jones potential was used to represent the interactions between all particles except for that between the chain elements forming a bond in the polymer chain, for which the interaction was expressed by a harmonic potential. The self-diffusion coefficient and velocity autocorrelation function (VACF) of a polymer were calculated at various chain lengths $N_p$, and various interaction strengths between solvent molecules and a polymer chain element. For self-diffusion coefficients D, the Einstein relation holds good; as chain length $N_p$ increases the D value decreases, and D also decreases as ${varepsilon}_{cs}$ (the interaction parameter between the chain element and solvent molecules) increases. The relaxation time of velocity autocorrelation decreases as ${varepsilon}_{cs}$ increases, and it is constant for various chain lengths. The diffusion coefficients in various conditions reveal that our systems are in a free draining limit as is well known from the behavior of low molecular weight polymers, this also agrees with the Kirkwood-Riesman theory.