Global warming has extensively transformed Arctic sea ice from continuous level-ice coverage to unconsolidated ice floe fields. Whilst the ice floes have a mixture of different sizes and their locations are randomly distributed, contemporary computational models lack effective methods to generate floe fields with such a natural pattern. This work introduces two original tools that can generate realistic ice floe fields for computational models. They are a sequential generator that sequentially handles ice floes one by one, and a genetic generator based upon a genetic algorithm. Demonstration of the tools is given, presenting samples of generating various shapes of floes and arbitrary mixtures of different shapes. Furthermore, an example is provided that combines the generated floe field with computational work modeling a ship transiting in ice floes. In addition, the source code of the tools is sharable to the public.