<p>The hybrid Jet Mixing (PJ) method, which integrates mechanical mixing with high-pressure fluid injection, is increasingly utilized for ground improvement. However, estimating the precise column diameter and Unconfined Compressive Strength (<i>UCS</i>) prior to execution typically requires costly and time-consuming field trials. Traditional prediction models largely rely on fluid kinetic energy, which often fails to capture the complex soil-binder interactions in highly cohesive environments. To address this limitation, this study leverages a multi-project field database from cohesive soil site to develop robust empirical models for predicting PJ column diameter and UCS. The proposed formulations overcome the limitations inherent in classical energy-based theories by explicitly integrating the nozzle standoff distance (<i>e</i>) alongside jet energy parameters to account for the physical mechanics of mixing. Furthermore, the UCS model introduces the concept of “clay poisoning” to quantitatively address the detrimental interference of the soil’s Plasticity Index on cement hydration. By incorporating a novel soil resistance index (<i>S</i>_<i>soil</i>) that mathematically combines undrained shear strength and Plasticity Index, the developed models demonstrate high statistical reliability for standard datasets. Crucially, a rigorous outlier analysis—highlighting a significant deviation in highly plastic/organic soils—is conducted to establish the strict domain of applicability for these empirical equations. Ultimately, these formulations extend current design methodologies, providing practitioners with reliable, cost-effective tools for optimizing preliminary designs, reducing the extent of required field verification, and determining cement dosage and column geometry based on fundamental soil inputs.</p>