This work extends the current SotA of Graph Gaussian Processes (GGPs) to dynamic graphs and asses the performance of the proposed evolving-Graph Gaussian Process (e-GGP) in two simulated tasks where deformable objects are represented as a graph that evolves over time.