You write them more or less the same way. Molecular orbitals are still easily classified as pi or sigma, although their energetic ordering is not always the same as in the homonuclear diatomics.
Also, just to be clear, the nature of atomic orbital interactions is no different in the heteronuclear diatomics compared to the homonuclear diatomics. Any atomic orbitals of the appropriate symmetry will interact, and as the atomic orbitals are essentially the same (albeit different energy) in all atoms, so their interactions with the orbitals of other atoms are also the same. What changes is the degree of interaction between the atomic orbitals because the degree of interaction between any two atomic orbitals depends on the energy difference between them.
Take the O2 molecule as any example. Generally we draw the MO diagram as the 1S AO of the first oxygen atom interacting with the 1S AO of the other oxygen atom to form the 1σ MO and 1σ* MO, the 2S AO of the first oxygen atom interacting with the 2S AO of the other oxygen atom to form the 2σ MO and 2σ* MO, and so on. In fact, since the 1S AO and 2S AO have the same symmetry, the 1S AO of the first oxygen atom also interacts with the 2S AO of the second oxygen atom (and vice-versa), such that the 1σ MO technically contains some 2S AO character. We usually neglect this when drawing MO diagrams because the interaction with 1S and 2S is so small (because they are energetically far apart).
Since there are in principle an infinite number of AOs around each atom, then in principle each MO draws from an infinite number of interactions between them. The only interactions that are strictly forbidden are those between AOs with incompatible symmetry (and even these get relaxed in higher formalisms). If we drew every symmetry interaction between AOs, the MO diagram would be noisy to the point of useless, so we usually only indicate the "major" AO contributors to each MO. In homonuclear diatomics, the most significant interactions are those between "like" AOs on each atom, so drawing MO diagrams is pretty simple. In heteronuclear diatomics, the energy matchings are not always as easy to discern because the two sets of AOs do not line up neatly. Nevertheless, even though MOs draw from more AO-AO interactions, their symmetry characters are the same as they are in the homonuclear diatomics, even if their energetic orderings and contributing AO interactions are more complex.