Forensic palynology has been important in criminal investigation since the 1950s and often provides evidence that is vital in identifying suspects and securing convictions. However, for such evidence to be used appropriately, it is necessary to understand the factors affecting taphonomic variability (i.e. the variability in the fate of pollen grains before they are found during forensic examination). Here, we test the relative amount of pollen retained on clothing after a period of simulated light or heavy wear based on pollen and fabric characteristics. We also test the efficiency of forensic laboratory protocols for retrieving pollen from fabrics for analysis. There was no statistically significant difference in retention of fresh or dried pollen on any fabric type. There was a substantial difference in pollen retention according to wear intensity, with considerably more pollen being retained after light wear than after heavy wear. Pollen from insect-pollinated species was retained at higher concentrations than pollen from wind-pollinated species. This pattern was consistent regardless of wear intensity but pollination type explained more of the variability in pollen retention after light wear. Fabric type was significantly related to pollen retention, but interacted strongly with plant species such that patterns were both complex and highly species-specific. The efficiency of removing pollen with the standard washing protocol differed substantially according to plant species, fabric type, and the interaction between these factors. The average efficiency was 67.7% but this ranged from 21% to 93%, demonstrating that previous assumptions on the reliability of the technique providing a representative sample for forensic use should be reviewed. This paper highlights the importance of understanding pollen and fabric characteristics when creating a pollen profile in criminal investigations and to ensure that evidence used in testimony is accurate and robust.