The problem of regional pole placement by static output feedback is studied for a class of linear descriptor systems that presents two caracteristics found in some physical models: regularity and absence of direct action of control inputs on the algebraic variables. Thus the structure of the mathematical model of this class of systems is exploited. We show, in particular, how some results from the classical theory of normal (differential) linear systems can be used to solve the considered problem. The paper provides a set of necessary and sufficient conditions to characterize the existence of an output feedback that places the finite closed-loop poles in particular regions of the complex plane. A technique based on the use of an orthogonal decomposition and on the solution of two-coupled Lyapunov equations is also proposed. A numerical example illustrates the approach.