For instance if I want to search for genes encoding particular enzymes which bioinformatic tool is better and how do I select the best genes to investigate?
You can do protein blast search and with the identified protein of interest you can go for MUSCLE alignment with already functional sequences of enzymes and look for the conserved catalytic active site residues. If your protein contains all the conserved residues you can go for cloning work or further experiments.
It will depend on which genome sequence you want to do the mining. If you are interested in searching a particular enzyme in the already published genomes then you can go, for example to:
http://www.genome.jp/kegg/annotation/enzyme.html
and do the search by name for the enzyme of interest. Then select one from the results and you will see under table/tab "Genes" all the genes encoding that enzyme across all of the species' genomes that encode it (click "show all" there and you will see the full list - warning: it can be large).
On the other hand, if your genome is not yet published and/or it is not yet annotated, then (as Mangal and Priya already said) you can perform a BLAST search with the sequence of your enzyme of interest in the target genome sequence. If you have the protein sequence of your enzyme and just want to do a quick look up on your genome sequence then you will need TBLASTN (search of protein sequence query over a translated nucleotide sequence). However, if this is the case, I would recommend you firstly to look for genes signatures in your genome sequence (with a gene-finder -- if prokaryotic then I would recommend prodigal, for example) and then perform the search using the protein sequences as target, which are a normal alternative output of any gene-finder (I mean, construct a protein sequence database with the protein sequences found by the gene-finder and then perform the search with BLASTP using your enzyme sequence as query).