Deficiency of succinate semialdehyde dehydrogenase (SSADH; aldehyde dehydrogenase 5a1 (ALDH5A1), OMIM 271980, 610045), the second enzyme of GABA degradation, represents a rare autosomal-recessively inherited disorder which manifests metabolically as gamma-hydroxybutyric aciduria. The neurological phenotype includes intellectual disability, autism spectrum, epilepsy and sleep and behavior disturbances. Approximately 70 variants have been reported in the ALDH5A1 gene, half of them being missense variants. In this study, 34 missense variants, of which 22 novel, were evaluated by in silico analyses using PolyPhen2 and SIFT prediction tools. Subsequently, the effect of these variants on SSADH activity was studied by transient overexpression in HEK293 cells. These studies showed severe enzymatic activity impairment for 27 out of 34 alleles, normal activity for one allele and a broad range of residual activities (25 to 74%) for six alleles. To better evaluate the alleles that showed residual activity above 25%, we generated an SSADH-deficient HEK293-Flp-In cell line using CRISPR-Cas9, in which these alleles were stably expressed. This model proved essential in the classification as deficient for one out of the seven studied alleles. For 8 out of 34 addressed alleles, there were discrepant results among the used prediction tools, and/or in correlating the results of the prediction tools with the functional data. In case of diagnostic urgency of missense alleles, we propose the use of the transient transfection model for confirmation of their effect on the SSADH catalytic function, since this model resulted in fast and robust functional characterization for the majority of the tested variants. In selected cases, stable transfections can be considered and may prove valuable.