We present results of density-functional tight-binding calculations on small titanium carbohedrene clusters (Ti metcars) Ti_mC_n with m = 7, 8 and n = 10–14. The structure optimization was performed within our newly developed genetic algorithm, which operates on a given (randomly chosen) population of clusters to generate new clusters of lower total energy by using certain evolutionary aspects. The structural and electronic properties of the resulting structures are analyzed. Especially Ti₈C₁₂, which is the most dominant and stable metcar in experiments, has been investigated, but did not turn out to be energetically favoured.