In this paper, a novel electromagnetic micropositioner is designed from an orthogonal 3-P(4S) parallel mechanism through the substitution method and modular design techniques. Preliminary prototype experiments show that the micropositioner possesses an excellent decoupling performance. Thus an independent control strategy is carried out for the motion control of the micropositioner. An RBF neural networks based adaptive backstepping terminal sliding mode controller is designed according to the nonlinearity characteristics of the actuator. Parameters of the system are identified with a genetic algorithm. Finally, the performances of the micropositioner and the developed control strategy are verified. Experimental results demonstrate that satisfactory performances can be achieved.