In this study we propose a novel formulation of a decision problem in R&D strategy. The problem is motivated by and applied to the context of technologies relevant to global climate change, but is characterized in general by an aggregate R&D decision-maker with a social welfare objective, technology diffusion markets subject to externalities in which private costs are minimized, and uncertainty in both technological and environmental factors. A technology strategy is defined as the allocation of R&D investment across several broad research programs, and the probability distribution over a discrete set of technological outcomes is modeled as a function of the technology strategy; thus the investment portfolio is the decision variable in the optimization. We also examine the relationship between the optimal technology strategy and market-based internalization policies such as emissions permits or taxes. The dynamic nature of the problem is emphasized in our methodology in order to properly account for several structural features, including the time lag between R&D investment and technology diffusion, the resolution of uncertainty over time, and the extended time horizon for climate change induced damages.