Before beginning your query, first make sure that you are using the correct database. There are currently five databases to choose from: blastn (for nucleotide sequences), blastp (for protein sequences), blastx (for nucleotide vs. protein translated sequences), blastn (for protein vs. nucleotide translated sequences), and blastx (for nucleotide vs. nucleotide translated sequences). For sequence identification, it is best to use either blastn or blastp (depending on whether you are using a nucleotide or protein sequence). Select the appropriate database and paste your sequence into the query sequence section on the search page. Before running your search, you can adjust the search parameters by organism or taxonomic group. You can also adjust the search program to narrow the speed and sensitivity of your search. For more information on these parameters, reference the BLAST How-to Guide here.
After initiating your search, the results are sorted by "best-match"—so your most similar result will be the first one listed. You can evaluate the level of similarity by looking at the scores to the left of each result description. The result with the highest "Max score" and "Total score" will be the most similar to your unknown sequence. The "e-value" is the error value: any e-value less than 1 x 10^-4 can be considered homologous to your unknown sequence.
If you wish to further evaluate similarities between your unknown and the result sequences, scroll down to "Alignments." Here, look at the "Identities" value. This will tell you how many amino acids in the result sequence exactly match those in your unknown sequence. Once again, the result with the highest number of matches will be the most similar.
This tutorial gives a great visual overview of navigating the query results for further help!