Assuming that you know exactly where in the full DNA strand the gene of interest is located, it is fairly easy to isolate the gene. There are several restriction enzymes you can use, it's just a matter of determining which one(s) are most appropriate for your gene of interest.
Remember that it is usually fine to have extra nucleotides on either end of the gene; the ribosomes will start and stop adding peptides based on start and stop codons, independent of how far the ribosome moves before reaching the start codon. Analyze your DNA strand and try to find a restriction enzyme that will slice the strand somewhere before the gene begins, and usually another that will slice it somewhere after it ends, although sometimes the same enzyme will work for both ends of the gene.
Edit: As @Untitpoi mentioned, if the restriction enzyme's slice sequence is found within the gene of interest, it will cut your gene in two (or more). It is very important to ensure that the enzyme(s)s you choose do not slice the gene of interest, otherwise the strand will be useless. It's possible to re-join the strands (which, thanks to sticky ends, isn't too difficult), but why take the extra effort when you don't have to?