I confirm what jimmc said. The Smith chart is just another way of showing the same data reported on the table. Only for the Smith, the 10 ohms normalization was chosen for readability: with a standard 50 ohm chart the curves would have been too close to the center-left part of the circle to be readable; this is universally done for power transistors.
As regards impedance matching, you should carefully read the notes below the table: Zsource is defined as the impedance looking into the matching network from the transistor gate; Zload is defined as the impedance looking into the matching network looking from the drain.
So, if e.g. your load is a standard 50 ohm antenna, you should design the load matching network adding elements while going backwards from the output, until you obtain the correct Zload. This Zload is inclusive of the inductor connecting the drain to the power supply, which is one of the elements.
Same from Zsource: start from the 50 ohm input and add parts until you see Zsource, to be "seen by the gate".
When this is done, the device will be optimally driven and loaded according to manufacturer suggestion at the frequency of choice. However, since there are many ways to accomplish this matching, impedances could be "strange" on other frequencies and the device may choose to oscillate... If you have no experience, you had better build something similar to the schematic suggested in the datasheet, adapting values.