In this paper we present the determination of ultratraces of cadmium ions in water by means of a minicolumn (MC) flow-through preconcentration system coupled with graphite furnace atomic absorption spectrometry. The core of the system is a lab-made ion imprinted magnetic organosilica nanocomposite which is employed as filler of the MC for the selective retention of the analyte. In this case superparamagnetic magnetite nanoparticles were coated with an amine-functionalized shell and ion imprinted with Cd(II) by a simple sol-gel co-condensation method. The setup was completed with the inclusion of a magnet fixed around the packed MC. This assembly - which is studied with an MII material for the first time here - allowed a homogeneous distribution of the solid on the walls of the MC, leaving a hole in the center and enabling the absence of material bleeding or obstructions to the free movement of fluids. Ion imprinted (MII) and non-imprinted (MNI) materials were studied for comparison purposes. Both were characterized and compared by DRX, FTIR, and SEM and their magnetic behavior by magnetization curves. Batch experiments showed an equilibration time of less than 10 minutes and a maximum adsorption pH of around 7 for both solids. The maximum capacity for MII was greater than that of MNI (200 mg g-1 and 30 mg g-1 respectively) and thus, the former was chosen for analytical purposes. Under MC dynamic conditions, sample and elution flow rates, volumes of the sample and eluant, and type and concentration of the most suitable eluant have been thoroughly investigated and optimized. Under the optimal experimental conditions, the MII filler showed a preconcentration factor of 200, a limit of detection of 0.64 ng L-1, a linear range of 2.5-100 ng L-1, RSD% of 1.9 (n = 6; 10 ng L-1) and a lifetime of more than 800 cycles of concentration-elution with no loss of sensitivity or need for refilling. The effect of potentially interfering ions on the percent recovery of cadmium was also studied. The proposed method was successfully applied to the determination of traces of Cd(II) in osmosis and tap water with recoveries of 98.0-101.3%. A comparison with similar methods is also provided.