Spectrometric oil analysis is of great importance for wear condition monitoring of gearbox.In this context,the contents of main elements compositions in the bench test of heavy vehicle gearbox are obtained by atomic emission spectrometric oil analysis first.Then correlation analysis of the test data and wearing mechanism analysis are carried out to get the metal element which could be used to describe the wearing and failure of the gearbox.The spectrometric data after filling/changing oil are corrected,and the laws of the contents of main elements compositions during tests are expressed as linear functions.After that,the reliability assessment is executed with considering the degradation law and discreteness of test data,in which the mean and standard deviation of normal distribution of spectrometric oil data at each time point are adopted.Finally,the influences of the threshold are discussed.It has been proved that the contents of metal element Cu,which is got by spectrometric oil analysis of different samples,could be used to assess the reliability of heavy vehicle gearbox.The reason is that the metal element Cu is closely related to the general wear state of gearbox,and is easy to be measured.When the threshold of Cu content is treated as a constant,bigger threshold means higher reliability at the same time,and the mean value of threshold has significant impact on the reliability assessment results as R>0.9.When the threshold is treated as a random variable,bigger dispersion of threshold means smaller slope of reliability against time,and also means lower reliability of gearbox as R>0.9 at the same time.In this study,the spectrometric oil analysis and probability statistics are used together for the reliability assessment of gear box,which extends the application range of spectrometric analysis.