Linear transformations of special relativity considered in [Albert Einstein, Zur Elektrodynamik der bewegte Körper. Annalen der Physik. 17 (1905) 891–921] are applied to Hamiltonian systems and Dirac equations, as presented in [Paul A.M. Dirac, Lectures on Quantum Mechanics, Belfer Graduate School of Science, New York, 1964]. To account for Weber’s electro-dynamic law of particle attraction and for the relativistic increase of the mass in particle accelerators, an extension of the second Newton’s law for motion subject to external forces that may depend on accelerations and higher order derivatives of velocities is considered and the Buquoy–Mestschersky generalization for motion of bodies with variable masses is included. The causality of systems driven by such forces is assured by consideration of left higher order derivatives in the right-hand sides of the equations of motion. The consistency condition is presented, and existence of solutions for the equations of motion driven by forces with left higher order derivatives is proved, leading to the generalized Lagrange and Hamilton equations that incorporate those extensions. Then, non-holonomic Hamiltonian systems with time-invariant constraints introduced by P.G. Bergmann and P.A.M. Dirac are considered, as suggested by Dirac for atomic models. Since one and the same process or particle may be observed as different images, the procedure is developed for identification of processes in moving systems by inverse relativistic transformations applied over small intervals of time to discrete experimental measurements obtained as the images of those processes in the observed (relativistic) coordinates. It is demonstrated that motions and processes evolving in still and/or moving systems can be described in the proper (of a still system) and/or relativistic (of different moving systems) coordinates in one common system of equations under the condition that all components of that system are referred to one and the same time of a still or moving observer. The results open new avenues for further research in relativistic systems theory and provide a basis for development of computing software for process identification in the images transmitted from distant or fast moving systems.