<p>Metal elements, critical to diverse industrial applications, also pose substantial environmental pollution risks when mismanaged. Conventional metal ion detection methods, however, are limited by high operational costs, bulky equipment and poor portability for on-site analysis, suggesting the need for an accurate, cost-effective and portable sensing strategy that is capable of addressing various metal ions directly from a heterogeneous sample. In this study, we engineered a <i>Mycobacterium smegmatis</i> porin A nanopore by incorporating an iminodiacetic acid ligand at its constriction site, creating a versatile sensor that is capable of simultaneously identifying ten divalent metal ions, namely, Sn²⁺, Cu²⁺, Pb²⁺, Cd²⁺, Mn²⁺, Zn²⁺, Fe²⁺, Co²⁺, Mg²⁺ and Ni²⁺. By integrating machine learning algorithms, this approach achieved a remarkable validation accuracy of 0.996. When applied to natural water samples, the strategy effectively demonstrated its practical utility for real-world environmental monitoring and metal ion detection.</p>