The energy source natural gas represents an attractive alternative fuel for SI engines due to its suitable chemical properties. But the lower performance yield and the insufficient infrastructure lead to a rejection by the customer. One approach to gain acceptance is the improvement of the driveability of the natural gas-powered vehicles. Direct injection of CNG is a technology with potential to achieve this objective. Nevertheless, the combustion concept has high requirements on injection system and mixture formation. Among other challenges the methane slip raises major issue. The compact molecular structure of the CNG's main component, methane, leads to an extremely slow reactivity, which expresses itself in a higher catalytic converter light-off temperature and thus aggravated exhaust aftertreatment. Especially, since methane's global warming potential is 21 times higher than CO 2 , the greenhouse gas footprint is getting worse again. Thus, for the future implementation of such a combustion system concept it will be necessary to compensate the limited conversion rates of the catalytic converter with low raw emissions. For this, the comprehension of the mixture formation is necessary. This paper deals with different injection strategies at part load and charged operation to achieve HC raw emissions as low as possible. The testbench experiments are executed with a two-cylinder SI engine under homogeneous stoichiometric conditions. The DI injector used for this study is a prototype model fabricated by Robert Bosch GmbH. In summary, the need of adaption of the injection strategy to the respective engine map area is becoming evident. Furthermore, various injection timings and injection splitting strategy turned out as an appropriate measure to make the most of CNG-DI's potential and to achieve low HC raw emissions.