In order to perform seismic inversion and have information about the subsurface, horizontal well-log data is normally used as constraints. These data contain very detailed information on a short scale. However, since the next step in the geo-steering of a borehole always depends on the previous one, the location errors accumulate and, as the drilling acquisition goes further and deeper, the uncertainty of the position and hence the borehole measurements also grows. To image the subsurface and estimate the errors involved in the data analysis process, we perform seismic inversion combining different sources of uncertain measurements, such as uncertain seismic data as well as uncertain well positions and well-log data. This paper presents a new method to incorporate these different sources of uncertainties and on different scales. We simulated various realizations of the subsurface and, taking into account the growing uncertain position of the well location, we could observe the variability that the errors create. As a result, the computed subsurface model and its error estimates are more realistic, and can better guide and optimise future drilling operations.