When to take the other body as point mass or apply parallel axis theorem on it?

If the point at which you want to measure angular momentum or consider mass moment of inertia, is not at the center of mass, you need to consider the parallel axis theorem.

Most generally, to measure the mass moment of inertia at a point A not at the center of mass C in 3D to the following

$$ mathrm{I}_A = mathrm{I}_C + m begin{vmatrix} y^2+z^2 & -x y & - x z - x y & x^2+z^2 & - y z - x z & - y z & x^2+y^2 end{vmatrix} $$

In shorthand notation the above is

$$ mathrm{I}_A = mathrm{I}_C - m [mathbf{r}times] [mathbf{r} times] $$

where $[mathbf{r}times] = pmatrix{0 & -z & yz & 0 & -x -y&x&0}$ is the 3×3 skew symmetric cross product matrix.

The same notation is used for the general rotation matrix about an axis $mathbf{z}$

$$mathrm{R} = 1 + sin theta [mathbf{z}times] + (1-cos theta) [mathbf{z}times] [mathbf{z}times] $$

The Parallel Axis Theorem does not have a range. It applies regardless of how far an object is from the axis of rotation.

However, as the distance $r$ from the axis of rotation increases, treating the distant object as a point particle becomes a better and better approximation. Whether you decide to use the Parallel Axes Theorem or treat the distant object as a point particle depends on what level of accuracy you want to obtain.

Suppose the moment of inertia through the centre of mass is $I_0=Mk^2$ where $k$ is the radius of gyration, which is on the same order of magnitude as the width/breadth/length $d$ of the object. Then the moment of inertia about a parallel axis at distance $r$ is
$$I=I_0+Mr^2=M(k^2+r^2)$$ If $r gg d approx k$ then $I approx Mr^2$ which is the moment of inertia of a point particle. The fractional level of accuracy you can expect from using the approximation is about $(frac{d}{r})^2$. So if $r approx 10d$ then the approximation will be accurate to about 1%.