In plants, an increasing interest for the comet assay was shown in the last decade. This versatile technique appears to be promising to detect the genotoxic effect of pollutants and to monitor the environment. However, the lack of a standardised protocol and the low throughput of the assay limit its use in plants. The aims of this paper are to identify key factors affecting comet assay performance and to improve its reliability and reproducibility. We examined the effect of varying several parameters on four different plant species: broad bean (Vicia faba), white clover (Trifolium repens), English ryegrass (Lolium perenne) and miscanthus (Miscanthus x giganteus). The influence of both internal (different nucleus isolation methods, presence or absence of filtration and lysis steps) and external (room temperature, light intensity) parameters were evaluated. Results clearly indicate that short chopping is more efficient to isolate nuclei than the standard slicing method. Filtration and lysis steps were shown to be unnecessary and thus should be skipped. Data also demonstrate that high room temperatures and light could induce DNA damage in isolated nuclei. Calibration tests with H2O2 or ethyl methanesulfonate revealed that a special attention should be paid to plant growing stage, leaf position and exposure duration.