Good example for your request http://doi.org/10.1007/s00894-016-2949-5

Short answer: no. Not from first principles.

Nodirali S. Normakhamatov's example is a successful example for the prediction of structures, but it's based on prior knowledge of geometries of the specific type of zinc complexes that form the core building block of the examined MOFs.

It's unfortunately still very extremely difficult to predict structures of completely new metal complex, as there is a multitude of factors that can play a role. There are projects underway that employ machine learning-based approaches to attempt to predict new structures based on the components, but as far as I know, there's still nothing for the end user. (I'm not even sure if there are already publications on that topic available.)

What usually is (still) done is making educated guesses for possible structures as input geometries, calculate those, and compare the resulting energies to decide which is the most likely outcome.

In most cases, you will have at least a general idea how your ligand could bind to your metal ion of choice, and based on this, you can think of possible geometries for your complexes. Depending on the flexibility of your ligands and the solvent you use while making the complexes experimentally, there still can arise a lot of problems, though.

it possible to predict complexation mechanism between organic molecule and metal via MOE and other software. There are some good letures on youtube on coordination complexes & Chemical their bonding.