In the past years, cellular metabolism of the immune system experienced a revival, as it has become clear that it is not merely responsible for the cellular energy supply, but also impacts on many signaling pathways and, thus, on diverse cellular functions. Label-free fluorescence lifetime imaging of the ubiquitous coenzymes NADH and NADPH (NAD(P)H-FLIM) makes it possible to monitor cellular metabolism in living cells and tissues and has already been applied to study metabolic changes both under physiologic and pathologic conditions. However, due to the complex distribution of NAD(P)H-dependent enzymes in cells, whose distribution continuously changes over time, a thorough interpretation of NAD(P)H-FLIM results, in particular, resolving the contribution of various enzymes to the overall metabolic activity, remains challenging. We developed a systematic framework based on angle similarities of the phase vectors and their length to analyze NAD(P)H-FLIM data of cells and tissues based on a generally valid reference system of highly abundant NAD(P)H-dependent enzymes in cells. By using our analysis framework, we retrieve information not only about the overall metabolic activity, i.e., the fraction of free to enzyme-bound NAD(P)H, but also identified the enzymes predominantly active within the sample at a certain time point with subcellular resolution. We verified the performance of the approach by applying NAD(P)H-FLIM on a stromal-like cell line and identified a different group of enzymes that were active in the cell nuclei as compared to the cytoplasm. As the systematic phasor-based analysis framework of label-free NAD(P)H-FLIM can be applied both in vitro and in vivo, it retains the unique power to enable dynamic enzyme-based metabolic investigations, at subcellular resolution, in genuine environments.