A novel magnetic and pH-responsive binary composite of calcium ferrite and multi-walled carbon nanotubes (CF-MWCNTs) as a drug carrier was synthesized via a two-step procedure, sol-gel auto combustion/ultrasonication, and characterized by various instrumentation techniques. Loading and release of metronidazole (MET) by the prepared composite were evaluated from various aspects. The study of the drug loading on the carrier was optimized by response surface methodology (RSM) and modeled by artificial neural network (ANN). The maximum drug loading efficiency of 98.21 % was achieved under RSM-based optimal conditions (0.06 g CF-MWCNTs, 30.50 mg L-1 MET, pH 5, and 15.23 min), and the ANN modeling revealed that the carrier's amount exhibited the key role in the loading process. The drug loading behavior was analyzed using the isotherm equations of Langmuir, Freundlich, and Temkin as well as Lagergren's pseudo-1st-order, Ho's pseudo-2nd-order, and Elovich kinetic models. It was found that the MET loading onto CF-MWCNTs was congruent with the Langmuir isotherm, indicating the physisorption nature of the drug on the carrier surface. Loading kinetics and thermodynamics also showed that the involved process was swift (k2 = 0.02 g mg-1 min -1), spontaneous (Delta G degrees < 0), and endothermic (Delta H<degrees> = 17.89 kJ mol-1). Evaluation of the MET release at different pHs indicated that neutral and slightly alkaline environments (7.4 and 8.5) were suitable to sustainable and prolonged liberation. Pharmacokinetics studies demonstrated that Korsmeyer-Peppas was the preferred model, implying the Fickian diffusion of the drug from the carrier. The MTT test disclosed that the CF-MWCNTs did not manifest any toxic effect toward A549 cell lines. Overall, the as-prepared CF-MWCNTs composite is a suitable platform for drug delivery applications according to its various merits.