Microbial whole-cell sensor has been widely used to assess bioavailability and risk of toxic elements, but their environmental use is still limited due to the presence of other interfering pollutants and the nonspecific binding in cells, which leads to inaccurate results. Here, we proposed a strategy combining Escherichia coli sensor set with binary regression models for the specific detection of bioavailable cadmium (Cd), lead (Pb), and arsenic (As) in a co-polluted environment. Initial tests suggested that the sensor set respectively termed pcadCluc, pzntRluc, and parsRluc could be classified into two groups according to their specific response to Cd, Pb, and As: group 1 (pcadCluc and pzntRluc) induced by a Cd-Pb mix and group 2 (parsRluc) induced by a Cd-As mix. Based on the variance in responses of each sensor to mixtures of target elements, three binary linear equations for two sensor groups were set up to calculate the individual concentrations in the mixture solutions. This method was then used to quantify the bioavailable Cd, Pb, and As in soils from a co-polluted mining region and to compare the results with other methods. Results showed that the conventional single target sensor method overestimated the bioavailability of each element, while sensor set was credible for accurate bioavailable Cd, Pb, and As quantification and comparable with the results from inductively coupled plasma mass spectrometry (ICP-MS) analysis. Our method can potentially be extended to cover the specific detection of other bioavailable toxic elements in different environmental settings.