In this study, positive-mode laser desorption-ionisation ultra-high-resolution mass spectrometry (LDI-FT-ICR-MS) was applied to study combustion aerosol samples obtained from a ship diesel engine as well as the feed fuel, used to operate the engine. Furthermore, particulate matter was sampled from the exhaust tube using an impactor and analysed directly from the impaction foil without sample treatment. From the high percentage of shared sum formula as well as similarities in the chemical spread of aerosol and heavy fuel oil, results indicate that the primary aerosol mainly consists of survived, unburned species from the feed fuel. The effect of pyrosynthesis could be observed and was slightly more pronounced for the CH-class compared to other compound classes, but in summary not dominant. Alkylation pattern as well as the aromaticity distribution, using the double bond equivalent, revealed a shift towards lower alkylation state for the aerosol. The alkylation pattern of the most dominant series revealed a higher correlation between different aerosol samples than between aerosol and feed samples. This was confirmed by cluster analysis. Overall, this study shows that LDI-FT-ICR-MS can be successfully applied for the analysis of combustion aerosol at the molecular level and that sum formula information can be used to identify chemical differences between aerosol and fuel as well as between different size fractions of the particulate matter.