Uncontrolled soil contamination by potentially toxic elements (PTEs) poses serious threats to environmental and public health in mining-intensive regions. Against this background, this study assessed the distribution, sources, ecological impact, and human health risks of eight PTEs (Cr, V, Zn, Pb, Ni, Cu, Co, As) using 1,884 soil samples collected across Tanzania’s Singida Region. Samples were analysed with Inductively Coupled Plasma Mass Spectrometry (ICP-MS). We applied a novel integrated framework combining Self-Organizing Maps (SOM), probabilistic human health risk assessment (HHRA), and fuzzy synthetic evaluation (EW-FSE) to improve spatial analysis and risk classification in under-studied regions. Mean PTE concentrations (mg/kg) were highest for Cr (62.55) and V (61.18), while Pb (25.32) and As (1.85), exceeding reference levels at numerous sites. Pollution indices revealed 59% of sites exceeded contamination thresholds, and 86.31% of samples surpassed the UCC for Pb. High coefficients of variation and extreme skewness for As indicated localized contamination from point sources. SOM analysis revealed two clusters; one geogenic (ultramafic/mafic lithologies) and one anthropogenic (linked to mining and waste). EW-FSE identified As (58.2%) as the major contributor to ecological risk, followed by Ni and Co. Probabilistic HHRA showed children are at higher risk, with Cr and Ni driving non-carcinogenic and carcinogenic hazards respectively. This integrated framework represents a novel contribution to regional-scale environmental geochemistry in sub-Saharan Africa. It addresses key gaps in source identification, spatial clustering, and uncertainty-based risk evaluation, and provides actionable insights for land-use planning, contamination control, and public health protection in mining-affected areas.