Fossil flowers are essential to infer past angiosperm evolutionary processes. The assignment of fossil flowers to extant clades has traditionally relied on morphological similarity and on apomorphies shared with extant taxa. The use of explicit phylogenetic analyses to establish their affinity has so far remained limited. In this study, we built a comprehensive framework to investigate the phylogenetic placement of 24 exceptionally preserved fossil flowers. For this, we assembled a new species-level dataset of 30 floral traits for 1201 extant species that were sampled to represent the stem and crown nodes of all angiosperm families. We explored multiple analytical approaches to integrate the fossils into the phylogeny, including different phylogenetic estimation methods, topological-constrained analyses, and a total evidence approach combining molecular and morphological data of extant and fossil species. Our results were widely consistent across approaches, with minor differences in the support of fossils at different phylogenetic positions. The placement of some fossils is in agreement with previously suggested relationships, but for others, a new placement is indicated. We also identified fossils that are well constrained within particular extant families, whereas others showed high phylogenetic uncertainty. Finally, we present recommendations for future total evidence analyses, regarding the selection of fossils and appropriate methodologies, and provide some perspectives on how to integrate fossils into the investigation of divergence times and the temporal evolution of morphological traits.