<p>Strontium hexaferrite (SrFe₁₂O₁₉) is a magnetic material with distinct features such as low production cost, excellent chemical stability, non-toxicity, resistance to corrosion and tunable coercivity. Magnetic induction coercivity of this compound contributes significantly to its wider applications in magnetic sensors, magnetic recording media, microwave devices, magnetic wave shielding and other biomedical applications. Experimental synthesis of strontium hexaferrite as well as introductions of different cations in the crystal structure at different concentrations through doping for magnetic induction coercivity enhancement, are laborious and costly. This necessitates the need for computational methods of coercivity prediction and enhancement. This work develops extreme learning machine (ELM) based model using ionic (Ionic) radii of the doping cations descriptors and crystal lattice (Latt) descriptors with sine (SN) and sigmoid (SG) activation functions. The predictive models were developed and validated using experimental data extracted from one hundred and twenty-two strontium hexaferrite compounds. Results of Ionic-ELM-SN, Latt-ELM-SG and Ionic-ELM-SG intelligent predictive models were compared with stepwise regression (STR) based model using root mean square error (RMSE), correlation coefficient (CC) and mean absolute error (MAE) metrics. Ionic-ELM-SN performs better than Latt-ELM-SG, Ionic-ELM-SG and Latt-STR predictive models with improvement of 27.06%, 35.61% and 60.49% respectively, using MAE metric on testing strontium hexaferrites. The developed predictive model was further validated using new experimental data and the obtained magnetic induction coercivity agreed excellently with the measured values. The developed predictive intelligent models would facilitate magnetic induction coercivity enhancement in strontium hexaferrite based compounds for several industrial and technological magnetic applications.</p>