To investigate the unregulated Ras activation associated with cancer, we developed and validated a mathematical model of Ras signaling. The model-based predictions and associated experiments help explain why only one of two classes of activating Ras point mutations with in vitro transformation potential is commonly found in cancers. Model-based analysis of these mutants uncovered a systems-level process that contributes to total Ras activation in cells. This predicted behavior was supported by experimental observations. We also used the model to identify a strategy in which a drug could cause stronger inhibition on the cancerous Ras network than on the wild-type network. This system-level analysis of the oncogenic Ras network provides new insights and potential therapeutic strategies.