<p>Pollution is the third most important trigger of extinctions globally, threatening biodiversity. Phytoremediation is a sustainable and cost-beneficial solution to restore soils, with a narrow, however, focus on either the plant host or the environment, overlooking the critical role of microbes within the complex plant-soil system where pollutant biodegradation and transformation occur. We ask here with a meta-analysis of 45 existing studies, what parameters determine the successful re-establishment of microbial communities in phytoremediated sites. Overall, phytoremediation increased richness metrics (OTUs, Chao1) in both bacterial (lnRR = 0.24, <i>p</i> &lt; 0.0001) and fungal (lnRR = 0.14, <i>p</i> = 0.02) communities; however, the diversity (Shannon and Simpson) was constrained by dominant taxa, which may reflect ecological pressures limiting recovery in contaminated soils. Bacterial communities responded rapidly to phytoremediation but exhibited short-term effects, compared to fungi. Phytoremediation outcomes varied by pollutant type, being stronger in soils with organic contaminants than with heavy metals. Finally, phytoremediation showed dependencies to the bioconcentration factor, plant genotype, and pH. To our knowledge, this is the first meta-analysis to quantify the role of plant functional traits in microbial responses during phytoremediation, with deep-rooted trees outperforming herbaceous plants. These findings guide the selection of plant species for effective phytoremediation.</p> Graphical abstract <p></p>