Manifold attempts have been made to produce reliable and efficient means of detecting problems in ships and give proper help in time. The VDR (Voyage Data Recorder) in ships are used to review events that took place in the moments before an incident and help to identify the cause of marine accidents. Voyage Data Recorder contains audio information about incidents occurring inside and outside the ship, mainly the captain’s room. Presently, In this VDR the data is saved only in RAM. Sometimes the data can be deleted by an unknown person or an officer, who is responsible for that incident. Since it is possible to delete data from VDR, the audio is immediately uploaded to server and instant audio processing is performed. The server extracts important features, analyses it immediately and stores it in the database. The important features are pitch, amplitude, positive amplitude, negative amplitude, zero crossing rate, and magnitude and convert audio time signal into frequency signal using Fast Fourier Transform(FFT). From the frequency, server extracts features such as mean, variance ,standard deviation, energy, skewness, and kurtosis. Finally, the system uses, thresholding and SVM classification to classify the signal, whether it is in normal class or abnormal class. The SVM classifier takes a set of input data and predicts. For each given input, two possible classes form the output. The SVM classification is already trained with possible training set. Now our new features are given to the SVM classification. If maximum number of data is in abnormal class, then immediately an alert is sent to the ship police and owner of the ship and also a report is generated, which is used to review the data and unwanted events that occurred in the ship for legal purposes.