Long-range research and development (R&D) planning must evaluate and compare programs with diverse objectives and uncertain outcomes, yielding multidimensional benefits and costs over many years. The relevance of several approaches in system theory to these considerations is discussed. Among those approaches are multiattribute utility analysis, decision analysis, time and state dependent preference analysis, portfolio selection, Paretian analysis, and interpretive structural modeling. A tentative procedure is outlined that integrates most of these approaches into one comprehensive methodology for R&D planning. National energy R&D strategic planning is taken as an example. In this methodology, a stochastic model of each R&D program is used to project its consequences in terms of improved future technology, as evaluated on several dimensions. A key feature of the approach is the evaluation of each new technology within the economic and technological context of the years of its deployment and use. Some aspects of implementation of the methodology are considered.