Environmental pollution of fossil fuels along with the rising cost of these finite energy resources are the major reasons for the accelerated growth of renewable energy sources especially wind and solar energy over the past few years. In the meantime, solar energy, an infinite source of clean energy, is on the rise as a response to the development of PV technologies and the improvement of PV panels efficiency. Significantpricereduction for PV panels has provided economic justification for a steep increase in the pace of large-scale PV plants installation over the last five years. Now in most industrialized countries, many utility-scale PV power projects are under construction or in operationat a scaleof at least 100 MWp. Traditionally, the main priority of the solar inverter control algorithms is the injection ofmaximum available active power from the solar arrays to the grid. More recently, due to the increment of inverters power rating in utility-scale PV plants the possibility of controlling reactive power has been provided which results in voltage stability margin improvement and powerlosses reduction. Therefore, independent power providers have shown interestin the capability of solar inverters to also control reactive power.There are many inverters in a large-scale PV power plant which their output power widely varies during a day. Especially in cloudy conditions or at night active power output of inverters is much less than their rated power capacity. By employing appropriate control methods, these surplus capacitiesof thesolarinverters in a PV power plant can be usedforreactive powerinjection to the grid. Even in emergency conditions, the amount of active power exchange can be reduced to compensate reactive power. This thesis proposes a dynamic model for a PV power system based on instantaneous active/reactive power components which facilitates control of reactive power in PV inverters. Since instantaneous active/reactive power components are independent of any qd reference frame, thus, the designed controller is inherently robust against the system dynamics. It is also shown that using power components instead of conventional qd current components as the system state variables in the proposed model, reduces the nonlinearities in system dynamicsequations. Moreover, this thesis introduces two control methods based on this model. The first method adopts the sequential loop closing technique and the second one uses feedback linearization technique to control theactiveandreactivepowerof PV power plant independently. The validity and performance of the proposed model and control approaches are investigated through both time-domain simulation of a 120 kW grid-connected PV system and a low-power laboratory prototype of a PV inverter. Close match between simulation and experimental results verifies the accuracy of the model. It also indicates that this model can successfully control theactiveandreactive power independently apart from the salient features of controller such as fastresponse, zero tracking error, and disturbancerejection capability. Keywords: Dynamic modelling, Control of grid-connected PV system, Control of instantaneous active and reactive power components, Multivariable control of PV system, Feedback linearization method.