The presentation deals with an experimental database of spontaneous-speech characteristics. The database is intended for a dialog system which have been developed in the Department of Computer Science at the University of West Bohemia within the framework of the Copernicus project. This system is assembled with a speech prosody module for processing the output of the acoustic phonetic module.. Proceeding that analysis the features obtained are evaluated using a neural network so that a type of the sentence can be determined. These results were verified in a number of experiments (1000 sentences). Having analysed the linguistic model we proposed further to extend the types of prosody characteristics and to create a database. Thus, besides the intonation analysis the prosody module also involves subroutines which can evaluate the pitch, both in a sentence and in a word, and the pause. The prosodic characteristics including the sentence are stored in the database and consequently exploited by the linguistic module as an additional information used for recognizing and understanding the spontaneous speech. Processing the characteristics by usual methods of statistics the database can also be used to generate answers in the dialog system. The module was implemented in the C language an supported by the ORACLE database. For the user interface the environment SQL is suggested.