Today, human activities appear to be the main source of soil contamination. Due to its industrial activities, human shapes the environment and radically modifies the balance around the activity sites in which he lives. The question of the management of these contaminated sites appears essential in term of environmental point of view, and especially for health reasons when people feed on contaminated gardens near factories. The integrated, multi-analytical approach of this study conducted in the former industrial city of Viviez (Aveyron, France) in the framework of the European project Intereg SUDOE "Soil Take Care" aims to first identify the speciation of interest contaminants (Zn, Pb, Cd, and/ or As), define the environmental risk related to storage to finally establish a link between the present of each bearing phases and the associated health risk. For this purpose, 11 points of interest were sampled in order to have data on both the gardens close to the installations but also on other spots such as a hiking path. For this purpose, physical-chemical analysis of all samples has been performed (pH, conductivity, CEC, grain size, % of carbonates and organic matter). The bearing phases has been highlighted by µXRD, SEM-EDS and by XANES(SOLEIL synchrotron) experiments. The environmental risk has been evidenced thank to the BCR extraction and the standardized EN-12457 leaching procedure. For the health risk assessment, two steps are required. First, the bioaccessibility procedure UBM, based onto operational chemical methodology mimicking soil ingestion and its stay in the gastrointestinal tract of the human body, was realized then, some scenario was chosen to assess the non-carcinogenic heath risk for children and adults Hazard Quotient (HQ) for Zn, Cd, Pb, As and carcinogenic risk for As. Whatever the type of samples studied, a contamination in Pb, Zn and As is effective. The main contamination is shown for gardens soils with values up to 1500 mg kg^-1, 4000 mg kg^-1, 150 mg kg^-1 in Pb, Zn, As, respectively. The nature of bearing phases is quite similar whatever the samples, and is constituted by micro to nanoparticles of franklinite, and polymetallic-iron oxides, and affined by coupling µXRF, SEM and XANES. The sequential BCR extractions showed a common behavior between zinc and cadmium, with a high proportion of easily mobilizable fraction (F1). For Pb, the reducible and oxidizable fraction are predominantly representing approximately 30% each. Arsenic is link to the residual fraction, indicating a long-term stability in soils. These results correlate with the UBM bioaccessibility results highlighting that zinc and cadmium reach about 60% of bioaccessibility. For Pb, the average is about 40%. XAS experiment were also done onto each phase after extraction (i.e. onto solid and also liquid extraction) and confirm the data allowing to determine and quantify the specific dissolution phases. With the chosen residential exposure scenario, the risk threshold (Hazard Quotient) is very often exceeded in the case of children for lead and arsenic. Regarding the hazard index (HI = ∑HQ), considering the bioaccessible, all except one of the samples present a non-carcinogenic risk for children. Carcinogenic risk for As have also been highlighted.