Next-generation sequencing (NGS) technologies now offer the opportunity to move from population genetics to population genomics, even in non-model species. In the present study, two NGS platforms, the Roche 454 FLX and the Illumina Genome Analyzer were used to discover Single Nucleotide Polymorphisms (SNPs) in the muscle transcriptome of the European hake (Merluccius merluccius), which represents one of the most important demersal fisheries in Europe. Transcriptome sequencing and de novo assembly into unique contigs, in silico SNP detection, and validation by high-throughput genotyping yielded a set of 395 polymorphic SNP loci. Of these, 381 nuclear SNPs were used to scan genomic variation in 19 Atlantic and Mediterranean populations of M. merluccius. As all SNPs reside in the functional (transcribed) portion of the hake genome, it is expected that a subset of sequence variants might be located in genomic regions potentially involved in local adaptation, conveying greater resolving power for the detection of spatial structuring on a large and/or local scale. Combining genome scans and correlation analysis between SNP allele frequencies and environmental factors, several candidate loci that are potentially under divergent selection were identified, mainly genes involved in muscles structural organization/function and protein synthesis regulation, or implicated in energy metabolism. Such “outlier” loci showed high performance in resolving population structure, unveiling distinct genetic clusters at multiple spatial scales, which possibly correspond to locally adapted populations. Following a precautionary approach, these results suggest the need to revise the current management strategies of European hake stocks, particularly within basins.