With increasing confidence in effectiveness and reliability in recent years, the cochlear implant (CI) has been proven as an indication for severe and profound hearing loss and has brought considerable gains to many implant recipients. It should be noted, however, that hearing with a CI cannot be equated with that of typical hearing, and those differences affect sustained auditory language comprehension (Hahne et al., 2012). Furthermore, current research supports the findings that auditory sentence comprehension in adult CI recipients is determined by processing semantic before syntactic information (Hahne et al., 2012; Hahne & Friederici, 2001; Friederici, Hahne & Saddy, 2002). These findings may be due to the lack of prominence of the syntactic elements of spoken language, and the fact that syntactic elements are not always necessary to encode a spoken message. Additionally it might be assumed that variations in the process of auditory sentence comprehension mainly concern processes that cannot be captured in a top-down manner. In contrast, the use of semantic information through the use of contextual knowledge may be much easier. The present work therefore examines, whether postlingually deaf or hearing-impaired patients focus more on semantic information to interpret linguistic utterances correctly by incorporating contextual information, or apply alternative processing strategies. For this purpose listening training has been developed, which requires both the identification of semantic and syntactic features in the set, and to match sonically fragmentary gap sentences. Furthermore, the empirical study is to contribute to the question of whether to strengthen compensatory strategies or to focus on deficits to overcome hearing limitations. As part of an exploratory cross-over design, 42 postlingually deafened cochlear implant users participated in all training programs. The patients were aged between 20 and 70 years and had been fitted with cochlear implants unilaterally or bilaterally for at least 2to a maximum of 9 months. They were compared within each group and to a control group who received no training, regarding their linguistic performance in audiometric tests. Results show that deviations in the process of sentence comprehension mainly concern the extracting of morpho-syntactic information. Furthermore, patients mostly benefit from auditory sentence training which allows the inclusion of top-down driven processes. In summary, all patients benefited from auditory training, but particularly positive performance changes could be found in application of the semantic training set. The success manifests itself in positive changes in auditory speech understanding in quiet and in noise.