I see two "issues," though I hesitate to call them mistakes:
You're explicitly setting Creature c's age as 89, and then rewriting that age with the uninitialized default (!) of Creature d. If this is what you intend to do, then fine, but at the very least you're wasting a few cycles to set a value that you intend to throw out later.
You're possibly violating the JavaBeans naming conventions.
To explain the latter point: a lot of Java libraries and frameworks (notably JSP) rely on JavaBeans to treat the object as a component. I haven't dived deeply into the actual mechanisms used, but from what I've read it relies on Introspection of the JavaBeans class to infer properties and the types of those properties. By overloading your setter setYearOfBirth() to accept both an int and a Creature, you could throw off the Introspection. (See here for a decent introduction to JavaBeans.)
This is not a big deal -- its entirely possible that you won't use this class as a JavaBean, and if you do it's trivial to refactor it so it works. But your teachers would probably prefer something a little cleaner, like the following:
class Creature {
private int yearOfBirth=10;
public void setYearOfBirth(int year) {
yearOfBirth = year;
}
int getYearOfBirth() {
return yearOfBirth;
}
public static void main(String args[])
{
Creature c = new Creature();
c.setYearOfBirth(89);
Creature d = new Creature();
c.setYearOfBirth(d.getYearOfBirth());
System.out.println(c.getYearOfBirth());
}
}
Now all of your access to yearOfBirth comes via the public getter methods, which helps encapsulation, and will prevent the code from breaking if your main method moves to another class. (As Greg D correctly pointed out.)
Also, this has the added benefit of making the intent of your code clear, which becomes more and more important when you start writing code for others to maintain and modify.