For over 30 years, sperm morphology assessment has been one of the most common tests in evaluation of fertility. This review examines the clinical relevance of sperm morphology assessment in the diagnosis of infertility and in assisted reproductive technology, as well as its analytical reliability. Publications on the pathophysiology, the analytical reliability of the test and its clinical relevance in diagnosis and in Assisted Reproductive Technology (ART) were evaluated. This review compared and discussed study methodologies and results, including patient characteristics, preparation, smear staining methods and classification systems. The assessment of the percentage of some abnormalities such as for example thin head, amorphous head, or bent or asymmetrical neck is of little clinical use, and their pathophysiology is not well explained as most are physiological traits. Some studies have highlighted correlations between the percentage of normal forms and functional sperm abnormalities, as well as correlations with ability to conceive in vivo and, in some situations, with the success of intra-uterine insemination (IUI) or conventional IVF. However, except in the case of some specific sperm defects (easy to detect with 99 or 100% of spermatozoa affected) and which are often linked to genetic disorders (globozoospermia, macrocephaly, decapitated sperm syndrome and fibrous sheath dysplasia), sperm morphology assessment has very poor sensitivity and specificity in the diagnosis of infertility. Moreover, there is very little evidence that indices of multiple sperm defects [sperm deformity index (SDI), teratozoospermia index (TZI), and multiple abnormalities index (MAI)] are relevant. Above all, many publications report a major lack of analytical reliability of this test, mainly in assessment of the details of sperm abnormalities. Many questions arise concerning how and when sperm morphology should be assessed, and how to interpret the thresholds of normal forms. Questions are raised on the real clinical impact of this test.