Estimating kinetic parameters in heterogeneous solid catalytic reaction networks is known to being a difficult task. This work aims at proposing a down-to-earth methodology to obtain kinetic parameters from numerical experiments. We present three techniques: a multivariable linear regression model, a stochastic metamodeling, and an optimized Kriging interpolator connected to a least-squares method. We consolidate the methodology in two different applications. The first one is a process with few components in two reactions from where it was possible to acquire the reaction rate equations that fitted literature data. The second one is a complex industrial reaction network. The results showed that even if the candidate proposed reaction rate equations do not fit the experiments, it is possible to construct a mathematical metamodel that conforms to the behavior of the components. Statistical tests showed that in both cases the proposed models successfully fit the experimental data.