To aid in debugging efforts to increase yield, titer, and productivity of engineered metabolic pathways, computational models are increasingly needed to predict how changes in experimentally manipulable variables such as enzyme expression map to changes in pathway flux. Here, an ordinary differential equation model is developed for a heterologous mevalonate pathway in E. coli using kinetic parameters culled from literature and enzyme concentrations derived from Selective Reaction Monitoring Mass Spectrometry (SRM-MS). To identify parameters most important to further experimental investigation, a global sensitivity analysis was performed, which pointed to amorphadiene synthase activity as the main limiting factor for amorphadiene production. Furthermore, the model predicted that in this local enzyme expression regime, the overall pathway flux is insensitive to farnesyl pyrophosphate (FPP)-mediated inhibition of mevalonate kinase, not supporting a hypothesis that had previously been posited to be limiting amorphadiene production. To test these predictions experimentally, two strains were constructed: (1) a strain containing a homologous mevalonate kinase with weaker feedback inhibition, and (2) a strain with greater amorphadiene synthase expression. The experimental results validate the qualitative model hypotheses and accurately match the predicted productivities for the two strains, particularly when an in vivo-derived kcat for amorphadiene synthase was substituted for the literature value. These results demonstrate the utility of using kinetic representations of engineered metabolic pathways parameterized with experimentally derived protein concentrations and enzyme kinetic constants to predict productivities and test hypotheses about engineering strategies.