We have used laser capture microdissection (LCM) and fluorescence microscopy to isolate genetically labeled neurons from the Drosophila melanogaster brain. From native thin sections, regions of interest could be analyzed with a spatial resolution better than 50 μm. To exploit the specificity of LCM for lipidomics, catapulted tissue patches were directly collected on a reversed phase column and analyzed using an on-column extraction (OCE) that was directly coupled with liquid chromatography-multistage mass spectrometry (LC-MS(n)). With this approach, more than 50 membrane lipids belonging to 9 classes were quantified in tissue regions equivalent to a sample amount of 50 cells. Using this method, the limit of quantitation and the extraction efficiency could be estimated enabling a reliable evaluation of acquired lipid profiles. The lipid profiles of cell body- and synapse-enriched regions of the Drosophila brain were determined and found to be distinct. We argue that this workflow represents a tremendous improvement for tissue lipidomics by integrating genetics, fluorescence microscopy, LCM and LC-MS(n).