We present a modification to Self-Consistent Charge Density Functional based Tight Binding (SCC-DFTB), which allows computation based on approximate atomic charges. We obtain these charges by means of a machine learning (ML) process, which combines a Coulomb model with a neural network. This allows us to avoid the Self-Consistent Charge (SCC) cycles in SCC-DFTB calculation, while keeping its accuracy. The main input of the model are the atomic positions characterized by a set of Atom-Centred Symmetry Functions (ACSF). The charge inference from our ML algorithm is as close as $10^{-2}$ unit of charge from the exact SCC solution. Our ML-DFTB approach provides a good approximation of the density matrix and of the energy and forces with only one single diagonalization. This is a significant computational saving with respect to the complete SCC algorithm, which allows us to investigate bigger ensemble of atoms. We show the quality of our approach in the case of charged Silicon Carbide (SiC) clusters. The ML-DFTB Potential Energy Surface (PES) mimics rather well the SCC-DFTB PES despite its simplicity. This allows us to obtain the same geometric structure ordering with respect to energy for small clusters. The dissociation barriers for ion emission are well reproduced, which opens the way to investigate ion field emission and charged cluster stability. The ML-DFTB approach is obviously not limited to charged clusters nor to SiC materials. It opens a new route to investigate larger clusters than investigated by standard SCC-DFTB, as well as surface and solid state chemistry at the atomic level.