Combining neural network based models with mechanistic models could potentially improve the prediction quality, reduce unexpected outcomes, help with lowering the training efforts and yield more interpretable, explainable and generalizable models. In our previous work, we introduced Systems Dynamics Aware Neural Networks (SDANNs) that facilitate incorporating known dynamics of the system into neural network training by following a very flexible programming pattern. SDANNs combine the neural networks and Systems Dynamics Modeling. Benefits include more accessible modeling, helping the practitioners when mathematical tools have limited applicability, and enabling modeling more complex behavior than physical systems. Systems dynamics aware neural networks approach opens possibilities in economics, finance, operational research, policy making, corporate strategies and other fields with complex problems. In this paper, we demonstrate how the parameters of complex systems can be estimated using Systems Dynamics Aware Neural Networks. We argue that the problem of parameter estimation of complex systems can be tackled by following the same flexible programming pattern in the training loop of the Systems Dynamics Aware Neural Networks.