We consider the case of methanol production in cold dark clouds, also known as quiescent cores, for which recent work shows that a purely gas-phase synthesis is unlikely to produce a sufficient amount to explain the observational fractional abundance of approximately 10(-9). Moreover, recent experiments appear to confirm a previous hypothesis that methanol can be formed on cold grain surfaces by the hydrogenation of CO via successive reactions with hydrogen atoms. In this paper we consider two ways of including the surface formation of methanol into chemical models of cold dark clouds. First, we use a gas-phase model and artificially include the surface formation of methanol in the same manner that the formation of molecular hydrogen is included. Secondly, we utilize a gas-grain code with a new mechanism for desorption following exothermic chemical reactions on grain surfaces. The latter method can reproduce the observed fractional abundance of gas-phase methanol and many other gas-phase species in the well-studied cold dark cloud TMC1-CP but the best fit to the observational data occurs at times significantly later than at ages estimated from gas-phase models.