The general idea of the research is based on the assumption that an oil well with an installed sucker rod pump (SRP) emits a characteristic sound spectrum that can be assessed. Every change to the system (wear, beginning of failures, etc) should be reflected in a corresponding change of the sound, creating thus a correlation. The scope of the research is to study noise, produced by a well and to stude relationship between emitted sound and a production state of the SRP. Correlation will be researched on the basis of dynamometer cards (DC) and actual production events. Noise represents a function of dynamic behavior of fluids, gas, downhole, and surface equipment. Noise created by this system is recorded in on-line mood with the help of a special device installed on the wellhead. The sound data then are transmitted, uploaded to a server, and available to handle. The acoustic analysis is based on the signal processing techniques combined with statistical tools. They are utilized in order to obtain sound features that are correlated to dynamometer cards and DC features. The result of the research supports application of the sound as a method for SRP supervising. High correlation coefficients corroborate relationship between SRP sound and DCs. In the research, the sound is applied to model the DC features for the SRP monitoring. This widens the application of acoustics as monitoring and SRP state prediction tool.