Abstract The effectiveness of two amendments for the in situ remediation of a Cd- and Ni-contaminated soil in the Louis Fargue long-term field experiment was assessed. In April 1995, one replicate plot (S1) was amended with 5% w/w of beringite (B), a coal fly ash (treatment S1+B), and a second plot with 1% w/w zerovalent-Fe iron grit (SS) (treatment S1+SS), with the aim of increasing metal sorption and attenuating metal impacts. Long-term responses of daily respiration rates, microbial biomass, bacterial species richness and the activities of key soil enzymes (acid and alkaline phosphatase, arylsulfatase, β-glucosidase, urease and protease activities) were studied in relation to soil metal extractability. Seven years after initial amendments, the labile fractions of Cd and Ni in both the S1+B and S1+SS soils were reduced to various extents depending on the metal and fractions considered. The soil microbial biomass and respiration rate were not affected by metal contamination and amendments in the S1+B and S1+SS soils, whereas the activity of different soil enzymes was restored. The SS treatment was more effective in reducing labile pools of Cd and Ni and led to a greater recovery of soil enzyme activities than the B treatment. Bacterial species richness in the S1 soil did not alter with either treatment. It was concluded that monitoring of the composition and activity of the soil microbial community is important in evaluating the effectiveness of soil remediation practices.