Magnetic field diagnostics of the transition region from the chromosphere to the corona faces us with the problem that one has to apply extreme UV spectro-polarimetry. While for coronal diagnostic techniques already exist through infrared coronagraphy above the limb and radio observations on the disk, for the transition region one has to investigate extreme UV observations. However, so far the success of such observations has been limited, but there are various projects to get spectro-polarimetric data in the extreme UV in the near future. Therefore it is timely to study the polarimetric signals we can expect for such observations through realistic forward modeling. We employ a 3D MHD forward model of the solar corona and synthesize the Stokes I and Stokes V profiles of C IV 1548 A. A signal well above 0.001 in Stokes V can be expected, even when integrating for several minutes in order to reach the required signal-to-noise ratio, despite the fact that the intensity in the model is rapidly changing (just as in observations). Often this variability of the intensity is used as an argument against transition region magnetic diagnostics which requires exposure times of minutes. However, the magnetic field is evolving much slower than the intensity, and thus when integrating in time the degree of (circular) polarization remains rather constant. Our study shows the feasibility to measure the transition region magnetic field, if a polarimetric accuracy on the order of 0.001 can be reached, which we can expect from planned instrumentation.