Don't you have a textbook or something that would have this outlined in it?
From HSC Online -
http://hsc.csu.edu.au/physics/core/space/9_2_1/921net.html#net2
"perform an investigation and gather information to determine a value for acceleration due to gravity using pendulum motion or computer assisted technology and identify reason for possible variations from the value 9.8 ms-2
* You may be performing an investigation that has been planned by your teacher. There are several suitable investigations that will achieve this purpose. One suitable investigation is described here.
A procedure for determining a value for acceleration due to gravity
A value for acceleration due to gravity can easily and accurately be measured by observing the motion of a pendulum.
1. Construct a pendulum at least one metre long, attached at its top to a support (such as a clamp connected to a retort stand) and with a small mass tied to its lower end to act as the pendulum bob.
2. Measure the length (l) of your pendulum, from its point of attachment to the centre of mass of its bob.
3. Pull the pendulum aside and release it so that it starts swinging. Using a stopwatch (or other device for measuring time), begin timing at an extreme of the pendulum’s motion and time ten full swings (one swing = back and forth) of the pendulum. Divide this time by ten to get a value for the average period (T) of the motion. Using this averaging technique tends to minimise random errors.
The period of a pendulum depends upon the length (l) and the value of acceleration due to gravity (g), as described in the following equation:
Equation for the period of a pendulum
Rearranging this equation gives an expression that can be used to calculate g.
Equation rearranged
4. Substitute your values for l and T into this equation to determine a value for g.
* As you gather information during your investigation, you may need to carry out repeat trials to confirm the reliability of your results. Also, you may want to use other, more accurate, timing devices or procedures to minimise the effect of random errors. "