A low-level distributed nonlinear controller for a DC MicroGrid integrated in a Smart Railway Station capable to recover trains' braking energy is introduced in this paper. The DC MicroGrid is composed by a number of elements: two different types of renewable energy sources (regenerative braking energy recovery from the trains and photovoltaic panels), two kinds of storages acting at different time scales (a battery and a supercapacitor), a DC load representing an aggregation of all loads in the MicroGrid, and the connection with the main AC grid. The nonlinear model of the MicroGrid is introduced, and a complete stability analysis is investigated to the purpose to meet power balance and grid voltage stability requirements. An Input-to-State Stability (ISS)-like Lyapunov function is obtained with a System-of-Systems approach, and it is utilized to develop the control laws for the converters in order to fulfill the dedicated objective each of them has. Simulation results, showing the desired grid behavior using the proposed nonlinear control laws, are introduced and compared with classical Proportional Integral (PI) linear controllers, with respect to performances and parametric robustness. The DC MicroGrid is shown to be able to operate braking energy recovery while performing load feeding and renewable energy integration and guaranteeing a proper DC voltage profile.