Astonishing ! Congrats !

A little technical question: if i am not mistaken, the coils in a Gottlieb EM are originally powered by 25 VAC.
You explain here above that you are using 24 VDC.
By rectifying 25 VAC, you get about 35 VDC.
The MOSFETs have a negligable voltage drop
So by what means do you produce 24 VDC ? Do you use an external SMPS ? Or do you use the recitified 25 VAC (so, 35 VDC) for the coils ?

If so, don't you fear to overdrive and overheat the coils ?

Yes, but that's not the point: modern games are equipped with coils designed to be driven by 50 - 70 V.
Your game has coils designed to work under 24 VAC.

Over a given period of time, the energy conveyed by DC is already higher than the energy conveyed by AC given the same voltage
Energy over time is power
So, driving a 25 VAC coil with 25 VDC will already produce a much higher power (both mechanical and heat dissipated).
But driving that very same coil with 35 VDC will definitevely put it under a lot of stress.

And not only the coil, but also the mechanics associated to it.
if you ever got a bent bumper leg which causes the bumber to remain stuck and quickly turn into a barbecue starter, you definitively will risk it now
And I'm not even talking about broken plastics

In this article, you will read an example of calculation on how to drive a coil designed for 120 VAC with DC.
The result is.... 34 VDC !
http://www.ecmweb.com/content/using-ac-coils-dc-power

I don't want to argue, but as you can see, I'm afraid you are taking some risks....

As about coils that will overheat, of course, it appears very rarely under normal conditions.
But you have to take into account the failure mode of your components.
Since MOSFET's failure mode is, in 90% of the cases turning dead short, your PWM is voided and your coil receives a full hit of the power supply
Furthermore, the High voltage rail may be sent back through the Mosfet's gate to the driver circuit, causing more failures in both the command circuitry (if not isolated through an opto-coupler) and causing a total collapse of the PWM circuitry.

If I were you, I would have an in-depth look on that awesome design of yours, just to be sure to not have it plagued by recurrent technical issues

You're assuming that nothing will ever fail.
I assume that, for example, the µC or anything in the interface between it and the Mosfet can fail or get stalled. Your PWM can be switched on 100% by a program bug
Even a short kick at 100% is strong enough to put the coil sleeve under mechanical stress since it was not designed for that
If your mosfet dies, the 35 VDC will be sent back through the gate to the driver circuit to your 5V rail and fry out just about everything on your board (unless the signal is fed through an opto-coupler to avoid that, but as you didn't mentionned that, i suppose it is not the case).

On the same subject, may I suggest you read https://hackaday.com/2018/05/23/mechanisms-solenoids/ , specially this part:
"While any solenoid will run on AC or DC, care needs to be taken to observe the coil’s specs. Solenoids represent an inductive load, and so their impedance is much higher in AC applications. So if a solenoid rated for 24 VAC were powered by 24 VDC, it would probably burn out quickly as the current through it would exceed the design specs. This could be avoided with a current limiting resistor or by lowering the DC supply voltage."
(Note that I personnaly disagree with the current limiting resistor suggestion)