Lignin is a polymer that greatly inhibits saccharification of plant feedstocks. Adjusting the composition or reducing the total lignin content have both been demonstrated to result in an increase in sugar yield from biomass. Yet, because lignin is essential for plant growth, it cannot be manipulated with impunity. Thus, it is important to understand the control of carbon flux towards lignin biosynthesis such that optimal modifications to it can be made precisely. Phenylalanine (Phe) is the common precursor for all lignin subunits and it is commonly accepted that all biosynthetic steps, spanning multiple subcellular compartments, are known, yet an <
em>
in vivo<
/em>
model of how flux towards lignin is controlled is lacking. To address this deficiency, we formulated and parameterized a kinetic model based on data from feeding <
em>
Arabidopsis thaliana<
/em>
basal lignifying stems with ring labeled [<
sup>
13<
/sup>
C<
sub>
6<
/sub>
]-Phe. Several candidate models were compared by an information theoretic approach to select the one that best matched the experimental observations. Here we present a dynamic model of phenylpropanoid metabolism across several subcellular compartments that describes the allocation of carbon towards lignin biosynthesis in wild-type Arabidopsis stems. Flux control coefficients for the enzymes in the pathway starting from arogenate dehydratase through 4-coumarate: CoA ligase were calculated and show that the plastidial cationic amino-acid transporter has the highest impact on flux.