To find a stationary point, you have to find where the graph is horizontal. The derivative is just the slope, so you have to find where this is zero.
Differentiating 3X^3 + 5X^2 + X -9 gives you 9x^2 + 10x + 1 so the stationary points are the roots of the quadratic equation 9x^2 + 10x + 1 = 0, which are x = –1/9 and x = –1.
So far so good, but you still don't know whether the stationary points are peaks (hills) or troughs (valleys). That's where the second derivative comes in. It gives you the slope of the first derivative.
Why do you need this? Well, at a peak, the second derivative is negative because the first derivative is changing from positive to negative as the original curve slopes up on the left then down on the right. At a trough, the reverse is true.
You find the second derivative simply by differentiating the first derivative. This gives 18x + 10. If you now plug in the values of the stationary points, –1/9 and –1, you get 8 (positive, indicating a trough) and –8 (negative, indicating a peak).
It's well worth the trouble to sketch the graphs of the original function together with its first and second derivatives. Then it will all make much more sense.