The sustainability of a course and training institute depends on the availability of students. There are many ways to promote the courses and training programs including promoting it through the institution's website. The visitor behavior of a website have hidden information that can be found using web usage mining approach. This study aims to discover the hidden information from the visitor patterns of course website. The data used are web access log data of August 2016. Web usage mining process was done using the Co-Occurence Map Sequential Pattern Mining using Bitmap Representation (CM-SPAM) algorithm which is available in the SPMF tool. Based on sequential pattern mining on the access log data, this study recommends improvements regarding the website structure and information that should be displayed on certain web pages. This study also found that the visitors of course website interested in three page types: one day seminar, tutorial and the training program.