In this paper, we consider the numerical optimal investment strategy of DC pension with time delay based on Markov chain approximation. A stochastic differential equation with time delay is obtained to describe the asset accumulation process. To maximize the utility function, an optimization problem is established. A novel numerical method is presented to solve the stochastic optimal control problem: First, a Markov chain is obtained, whose local properties are consistent with the stochastic differential equation with time delay. Then, based on the transition probability of Markov chain, the iteration equations of optimal control and value function are constructed, and the convergence analysis is given, too. Finally, numerical simulation is given, the results show that the numerical optimal control algorithm is feasible, and time delay has an impact on the retirement life of the policyholder.