Methods of intuition take heaps of practice. I started sketching curves on GeoGebra and in my head randomly to develop the intuition needed to just "SEE" how exactly the asymptote is approached, i.e. from above or below.
Techniques I employ include splitting each region between vertical asymptotes up, and observing the presence of both x-intercepts and stationary points. Horizontal/Oblique asymptotes are used for any region not bound between two vertical asymptotes.
E.g. for something like y=1/(x^2-1), I know about the stationary point at x=0, and I also know that between the vertical asymptotes x=-1, x=1, there are no x-intercepts. So the stationary point (which happens to be a local min) is going to, in a way, deflect the curve back down, so that both asymptotes are approached from below.
I also note the general shape of some things. Quadratic/Quadratic and constant/quadratic look the same, but linear/quadratic looks different.
Again, HEAPS of practice needed to develop intuition.