A Message is an information exchanged for personal or business purposes. These messages are generally targeted by spammers, resulting in forfeits in financial or monetary. Spam messages have grown significantly in different fields. Various machine learning based techniques have been used in the past for the detection of spam. A very few review works are available on spam detection techniques in the field of SMS, Email, Twitter and Online reviews. However, these studies have limitations of study of limited techniques from machine learning fields only. Also, an in-depth evaluation of performance for each of the suggested techniques are missing. In this paper, a detailed review of spam message detection techniques in five domains - SMS, Email, Twitter, Instagram and Online Reviews is done. Based on the reviews of state-of-the-art in the five domains, a generalized model for spam message detection is perceived and presented. Additionally, this paper provides a thorough review of the past researching the domain and detailed analysis is presented. The paper concluded with the future trends which can be used for message spam detection in near future.