Magnetic biosensors detect biological targets through sensing the stray field of magnetic beads which label the targets. Commonly, magnetic biosensors employ the “sandwich” method to immobilize biological targets, i.e., the targets are sandwiched between a bio-functionalized sensor surface and bio-functionalized magnetic beads. This method has been used very successfully in different application, but its execution requires a rather elaborate procedure including several washing and incubation steps. This dissertation investigates a new magnetic biosensor concept, which enables a simple and effective detection of biological targets. The biosensor takes advantage of the size difference between bare magnetic beads and compounds of magnetic beads and biological targets. First, the detection of super-paramagnetic beads via magnetic tunnel junction (MTJ) sensors is implemented. Frequency modulation is used to enhance the signal-to-noise ratio, enabling the detection of a single magnetic bead. Second, the concept of the magnetic biosensor is investigated theoretically. The biosensor consists of an MTJ sensor, which detects the stray field of magnetic beads inside of a trap on top of the MTJ. A microwire between the trap and the MTJ is used to attract magnetic beads to the trapping well by applying a current to it. The MTJ sensor’s output depends on the number of beads inside the trap. If biological targets are in the sample solution, the beads will form bead compounds consisting of beads linked to the biological targets. Since bead compounds are larger than bare beads, the number of beads inside the trapping well will depend on the presence of biological targets. Hence, the output of the MTJ sensor will depend on the biological targets. The dependences of sensor signals on the sizes of the MTJ sensor, magnetic beads and biological targets are studied to find the optimum constellations for the detection of specific biological targets. The optimization is demonstrated for the detection of E. coli, and similar optimization processes can be performed for the detection of other biological targets. Third, we demonstrate the new magnetic biosensor concept using a mechanical trap capable of detecting nucleic acids via the size difference between bare magnetic beads and bead compounds. The bead compounds are formed through linking nonmagnetic beads of 1 µm in diameter and magnetic beads of 2.8 µm in diameter by the target nucleic acids. The purpose of the nonmagnetic beads is to increase the size of the compounds, since the nucleic acid is very small compared to the magnetic beads. Alternatively, smaller magnetic beads could be used but their detection would be more challenging. Finally, an enhanced version of the magnetic biosensor concept is developed using an electromagnetic trap for the detection of E. coli. The trap is formed by a current-carrying microwire that attracts magnetic beads into a virtual sensing space. As in the case of the mechanical trap, the sensor signal depends on the number of beads inside of the sensing space. The distance which magnetic beads can be detected from by the MTJ sensor defines the sensing space. The results showed that the output signal depends on the concentration of E. coli in the sample solution and that individual E. coli bacterium inside the sensing space could be detected using super-paramagnetic beads that are 2.8 µm in diameter. In summary, this dissertation investigates a new magnetic biosensor concept, which detects biological targets via the size difference between bare magnetic beads and compounds of magnetic beads and biological targets. The new method is extremely simple and enables the detection of biological targets in two simple steps and within a short time. The concept is demonstrated for the detection of nucleic acid and E. coli.