In this paper, we develop an efficient algorithm to value options under discrete-time GARCH processes. We propose a procedure based on dynamic programming coupled with piecewise polynomial approximation to compute the value of a given option, at all observation dates and levels of the state...