I would say, it depends on the purpose of your research and analysis. However, partial least squares (PLS) is an alternative techniques. While SEM generally tests models of predicted relationships among observed and unobserved variables, PLS offers the same even if sample size is very small. Despite the similarities, both of the techniques contain some uniqueness...you will get an overview in order to make choice from the attached article.

Hasanuzzaman Tushar Thank you sir. Bujte parsi.