In concentrated solar power (CSP) stations, large arrays of mirrors which are capable of changing its orientation are used to reflect the incident solar energy to a stationary receiver kept at a distance. Such mirrors are often called as heliostats. The receiver contains a heat absorbing medium like molten salt. By absorbing the thermal energy reflected from thousands of heliostats, the temperature would reach around 6000C and the heat can be used in thermal power plants to generate steam and thus run a turbine to produce electricity. One of the biggest advantages of CSP over conventional energy harvesting from Sun is that it can generate electricity during night for long hours of time from the thermal energy stored during daytime. This eliminates the usage of batteries or any other energy storing methods. The conversion efficiency is also high in CSP due to the high temperature achieved. With prior knowledge of the station coordinates, viz., the latitude and longitude, the day of the year and time, the direction or the path of sun can be fully determined. Typically, the sun's motion is tracked by the azimuth-elevation (Az-El) or the target-aligned configuration heliostats. In both these approaches, the mirror needs to be moved about two axes independently using two actuators in series with the mirror effectively mounted at a single point at the centre. This arrangement causes the mirror to deform in presence of gusty winds in a solar field which results in loss of pointing accuracy. Typically a beam error of less than 2-3 mrad is desirable in a large solar field and this value also includes other sources of loss of pointing accuracy like gravity and wind loading. In order to prevent this, a rigid support frame is required for each of the heliostats. In this work, two three degree-of-freedom parallel manipulators, viz., the 3-UPU wrist and 3-RPS, have been proposed to track the sun in central receiver systems. The main reasons for choosing a parallel manipulator as heliostat are its desirable characteristics like large load carrying capacity, high accuracy in positioning the mirror and easy to obtain the inverse kinematics and convenient for real time control. The proposed parallel manipulators support the load of the mirror, structure and wind loading at three points resulting in less deflection and thus a much larger mirror can be moved with the required tracking accuracy and without increasing the weight of the support structure. The algorithm for sun tracking is developed, extensive simulation study with respect to actuations required, variation of joint angles, spillage loss and leg intersection has been carried out. Using FEA, it is shown that for same sized mirror, wind loading of 22 m/s and maximum deflection requirement (2 mrad), the weight of the support structure is between 15% and 60% less with the parallel manipulators when compared to azimuth-elevation or the target-aligned configurations. A comprehensive study on stroke minimization of prismatic joints is carried out. It is found that a stroke of 700 mm is required for a 2 m x 2 m heliostat at Bangalore when the farthest heliostat is at a distance of 300 m from the tower. Although, there is an extra motor required to track the sun, the 3-RPS manipulator is better than the conventional methods if the mirror area per actuator criteria is taken into consideration. Prototypes of the Az-El and 3-RPS heliostats were made with a mirror size of 1 m x 1 m. A PID controller implemented using MATLAB-Simulink and a low cost, custom made motor driver circuit is used to control the motion of the 3-RPS heliostat. The algorithm developed is tested on the prototype by tracking a point marked on the wall of the lab space and is found to have a tracking error of only 7.1 mrad. Finally, the actual sun tracking is carried out on the roof of a building reflecting the sun-light to a wall situated 6.72 m above and a distance of 15.87 m from the heliostats. The images are captured at various instances of time from 11:30 a.m. to 3:30 p.m. on October 15th and November 10th, 2016, tracking errors are quantified and it is demonstrated that the proposed 3-RPS parallel manipulator can indeed work as a heliostat in concentrated solar power plants.