<p>The global demand for borate minerals has increased substantially. However, efficiently targeting potential borate deposits remains a major challenge due to inherent uncertainties of subsurface characteristics. This study proposes an innovative workflow to quantify the uncertainty of constructed models for mineral explorations using the mixed Lp-norm magnetic inversion technique. The uniqueness of this workflow is interpreting the inverted subsurface within a logarithmic-probabilistic space, where the background and targeting anomalous features can be separated effectively. To be specific, the random sampling of user-defined parameters in mixed Lp-norm inversions accounts for the uncertainty, which generates diverse susceptibility models characterized by distinct spatial features. Statistical characteristics of these susceptibility values, constrained by rock samples, are subsequently calculated for the probabilistic interpretation of the inverted models. We applied the proposed workflow to a field case study focusing on the borate resource exploration in Liaoning Province, Northeast China. The resulting probabilistic models demonstrate a strong consistency with the observed drillhole data. It also supports the decision-making in determining the maximum depth of the drilling deployment in the survey area. The field results indicate a marked improvement in discovering the borate deposits using geophysical inversion technique, highlighting the potential of this workflow to enhance the efficiency of future exploration efforts.</p>