No, I do not expect that any book describes this procedure, and if one does it, it is not a good book.
The direct conversion from a formula to Matlab code is a very tiny part of the standard work during programming. Writing a computer program requires many other details to take into account, that you do not operate on symbolic expressions, but work on a physical machine with a limited numerical precision. A very simple example is this mathematical expression:
c = a + b;
It looks such trivial to convert it to Matlab:
c = a + b;
This will even work in many cases. But see this:
a = realmax('double'); % Bigest double value, about 1.7977e+308
b = 17;
c = a + b;
Slightly more complicated, but still looking trivial:
d = a + b + c;
Now test it with real values:
a = 1e17;
b = 1;
c = -1e17;
d = a + b + c % Replies 0 instead of 1!
You can imagine, that the problems will be much larger, if you want to solve an eigenvalue problem of a 10'000 x 10'000 matrix, find poles of a function, determine the gradient of a high-dimensional function, or integrate a stiff ODE with several force terms vibrating with very different frequencies.
You see, that "write equations in Matlab" does not hit the point. For solving problems books about numerics are useful. Because the same problems appear in any programmibng language, such numeric books need not focus on Matlab directly. A "good" book should concern the field you are working in, e.g. it matters if you are an image analyst of simulate multi-body dynamics. In a next step learning good programming patterns is useful for any programming language. And if these two topics are solved, reading the Getting Started chapters of Matlab's documentation should clear 98% of the questions concerning the formulation of mathematical problems in Matlab.
