It is more a process of ruling out the distributions your sample is not from. Start by graphing the data - boxplot, as a guide for outliers, skewness or symmetry, histogram (for a large enough sample), probability plots, etc.
Be aware that it can take very large sample sizes to determine whether (as an example) your data is better described by a normal distribution, a t-distribution, or some other symmetric distribution.
with what source of data are you working?