Nowadays, the system identification methods applied to civil structures are rising in order to get a better understanding of structural behavior and improve traditional analytical analysis. Accurate identification of the modal parameters of a structure is essential because it allows building a proper analytical model, and it discloses the difficulties that may not have been considered in analytical studies, as well as finding out the existence of structural damages or deterioration, and sometimes estimating the remaining life of the structure. A clear disadvantage of most experimental methodologies is to require of a long sampling time window that stresses the structure under test. This paper shows the effectiveness of a novel methodology based on the multiple signal classification (MUSIC) algorithm and its high-resolution properties, applied for identifying most of the natural modes and analyzing vibration signals in a truss-type structure by using a reduced sample data set and short sampling time window. It has the advantage of submitting the structure to a reduced fatigue and stress during testing as a difference from other works, where the analysis involves putting the structures under severe fatigue and stress. Identifying most of the natural modes in the truss-type structure is realized at first by locating the fundamental mode in a frequency region, and the other natural modes are identified in higher frequencies, where each of these natural modes is located in different frequency regions. Thus, the MUSIC algorithm can identify most of the natural modes in different frequency regions of a vibration signal successfully.