Printed circuit boards (PCBs) are composite laminate boards used widely in electronic industry. They facilitate the mounting of electronic components on them as well as to make electrical connections in them. Electronic cooling i.e. removal of heat from electronic components is a challenging task. The heat removal rate from the components also depends upon the thermal conductivity of the board over which it is fixed as it is a critical parameter. The challenge which it poses to the thermal engineers is the determination of thermal conductivity of these boards which are inherently anisotropic in nature. The present study aims to estimate the effective in-plane principal thermal conductivities (kxx and kyy) of anisotropic printed circuit boards by solving an inverse heat conduction problem with measured temperatures as input. Therefore, the present methodology involves three major parts namely the direct or forward problem solution, steady state experiments for obtaining temperature measurements and inverse problem solution. The direct problem involves the solution of heat conduction equation along with appropriate boundary conditions and is solved using commercial CFD package, Ansys Fluent. The experiments are done inside a vacuum chamber and steady state temperature measurements are obtained at 18(9 are used for estimation due to symmetry) preassigned locations. Inverse solution methodology employed using artificial neural network (ANN) is used to solve the inverse problem. The methodology is validated with an in-plane isotropic copper clad board by retrieving its in-plane thermal conductivity successfully. Following this, the inverse methodology is employed to estimate the in-plane thermal conductivities of four PCBs which differ in copper trace width and spacing.