A computational study of laminar and steady heat transfer is carried out with copper (Cu)-water nanofluid inside a lid driven cavity. Different conducting obstacles are placed at the optimum position for which the maximum thermal transport occurs. The thermal performance of the Cu-H2O nanofluid is found out at this optimum position with three different cylinder geometries (square, rectangular and circular), two Richardson numbers (0.01 and 1), and three volume concentrations of copper nanoparticles (0 %, 3 % and 5 %) in water. The Prandtl and Grashof numbers are considered as 6.2 and 10 4 , respectively. The fluid is heated by placing a differential heater at the left wall. All the walls except the upper one are in a stationary condition. The top moving wall, stationary bottom wall and the remaining portion of the left wall, where there is no heater, are made insulated. The rectangular cylinder is placed at two different orientations (vertical and horizontal). The results show that the shape of the cylinder contributes to a compelling role in efficient heat transfer. For both Richardson numbers, the Nusselt number is maximum with the case where the square-shaped cylinder is placed at the top left corner. The Nusselt number increases with an increase in the percentage volume of nanoparticles and reduces with an increase in the Richardson number.