Biologists have been investigating the plant defense response to virus infections for a long time. Nevertheless, its model has still not been developed. One of the reasons is the deficiency in numerical kinetic data that brings up the importance of the expert knowledge. Therefore, we based our work on acquiring domain knowledge of biological pathways which provided a basis for the construction of a dynamic mathematical model. The goal of our work was to model the major pathway of the plant defense response - the salicylic acid pathway - and determine its dynamic parameters that are in correspondence with the knowledge acquired from the biology experts. For this purpose, we first selected the Hybrid Functional Petri Net formalism to represent the model due to its intuitive graph representation important for the biologists and its mathematical capabilities necessary for the simulation. The salicylic acid model was manually constructed and curated. In addition, the knowledge related to the model variables was acquired from the biology scientists and formalized in the form of constraints. This enabled an automatic optimization search for the model parameters that violate the minimal number of constraints. If the simulation results do not match the expert expectations, the network structure and the constraint definition are revised and the optimization parameter search is repeated. The final results of our system are both simulation results and optimized model parameters, which provide an insight into the biological system. Our constraint-driven optimization approach allows for an efficient exploration of the dynamic behavior of the biological models and, at the same time, increases their reliability.