Both climate change and technical change are uncertain. In this paper we show the importance of including both uncertainties when modeling for policy analysis. We then develop an approach for incorporating uncertainty of technical change into climate change policy analysis. We define and demonstrate a protocol for bottom-up expert assessments about prospects for technologies. We then describe a method for using such assessments to derive a probability distribution over future abatement curves, and to estimate random return functions for technological investment in different areas. Finally, we show how these analytic results could be used in a variety of energy-economic models for policy analysis.