Motion behavior, Fault detection and failure recovery in a novel 6-DoF parallel manipulator is studied in this thesis. All possible faults in the manipulator are investigated and appropriate methods to detect and tolerate these faults are presented. Some of the proposed failure recovery methods are implemented on a real system. The investigated 6-DoF parallel manipulator is consisted of a fixed base, a moving platform, six upper arms and six lower arms; in which the upper arms connects to the lower arms and the moving platform with spherical joints at both ends. The lower arms are connected to the servo motor shaft by simple revolute joints. So the robot manipulator is consisted of six legs, which each leg contains two spherical joints, one revolute joint and two links, except for the base and moving platform parts. Kinematics equations of the system are generated both in the algebraic and differential form, considering the motion constraints. In the modeling process of the manipulator the joints flexibility and friction are ignored. It is well known that for these types of parallel manipulators, an exact analytic solution of the forward kinematics cannot be obtained. Therefor the forward kinematic equations of the manipulator are solved numerically using a Matlab/Simulink model. An ADAMS model was also developed to analyze the motion behavior of the system for the sake of comparison. This model was used to calculate the numerical results of both forward and inverse kinematics. Modeling of the behavior of the robot when it faces some sort of failure is the next step of the thesis. In this part three major possible faults in manipulators including link fracture, active joint jam and active joint force loss are studied. According to the lack of mechanical redundancy in this 6-DoF parallel manipulator, among the mentioned failures only the active joint jam failure is compensable. To do this, the change in DoF of the manipulator right after the failure occurrence is investigated firstly. Then two strategies, task prioritization and minimum velocity error, are proposed to compensate the failure consequences. To verify the simulation results of the two proposed strategies, the simulations are implemented on the real system. For this purpose, first the motor which has to encounter the failure is selected. Then the new motor inputs to compensate the effects of failure are calculated for both proposed strategies. The newly calculated motor inputs are then fed to the servo drives of the system. Servo drive transmits the appropriate signal to the servo motors in order to make them run. The encoders placed on the servo motors return the value of the servo motors angular position in each step. These values are compared with the firstly generated inputs to check ability of the proposed strategies in compensation of the failures. Keywords: 6-DoF parallel manipulator, Motion simulator, Rotary actuators, Failure analysis, Fault detection, Motion Recovery