The applicability of phytoextraction to remediate soils contaminated with heavy metals (HMs) depends on, amongst others, the duration before remediation is completed. The impact of changes in the HM content in soil occurring during remediation on plant uptake has to be considered in order to obtain a reliable estimate of the phytoextraction duration. To simulate the decrease in the HM content in soil and to assess the resulting decrease in the uptake of HMs by plants, contaminated soil was mixed with uncontaminated, but otherwise similar soil. Uptake of Cd, Pb, and Zn by the indicator plant Lupinus hartwegii and the Zn hyperaccumulator Thlaspi caerulescens (La Calamine ecotype) was a log-linear function of the in-situ measured HM soil solution concentrations. Over a wide range in dissolved Cd and Zn concentrations, uptake of these HMs by T. caerulescens was (much) greater than by L. hartwegii. Experimentally derived regression models describing the relationships between soil, soil solution, and plant were implemented in a HM mass balance model used to obtain estimates of the phytoextraction duration. For our target soils, estimates of the Cd phytoextraction duration using L. hartwegii or T. caerulescens increased significantly by more than 100 or 50 years when experimental soil-soil solution-plant relationships were used instead of the assumption of constant plant uptake of Cd. The two approaches gave similar results for phytoextraction of Zn by T. caerulescens.