The enhancement and improvement of the receptive textual competence for translation requires, firs of all, to identify the textual knowledge the translator has and to be able to structure it, and, secondly, to carry out tasks focused on the textual levels of his/her field of work. Acquisition of well-organised textual models constitutes a great advantage in the learning process of the translating competence, because it means to pass from intuitive to structured textual knowledge, and this has a positive impact both in the cognitive and the attitudinal sides of the competence. A full and clearly structured description of the text allows in practice a dissection of the field or work, thus facilitating the analysis of the source text (as well as that of the draft translation). Two types of training tasks, based on schemes of different dimension, are derived from that assumption: a) global schemes or representations of the full textual map: the description of all text levels and their internal connections, and b) partial schemes, or description of the text sequences.