What part of normalization do you not understand? What is the level of your database knowledge?
You are trying to organise your data structure (e.g. the tables and fields) such that redundancy can be eliminated (e.g. not storing the same data twice, not duplicating fields to represent the same thing multiple times, not storing unrelated data in the same table, etc). First remove any duplicate fields in a table (e.g. telephone1, telephone2, telephone3) to a new table such that there is only one field in that new table and rows it contains will represent the multiple instances [1NF]. Then make all tables cohesive; that is, making each table represent a unique construct, then add a primary key (a field that uniquely identifies each record in the table), and foreign keys (using the primary key of another table as a field to show the records are related) such that there are no relations which are many-to-many [2NF]. Then remove any fields in a given table are not part of the construct and so are not dependent on the primary key, and therefore not defined by the primary key [3NF].
I hope this helps, but is difficult to explain without examples as once you're used to it, it becomes very intuitive.