A Java programming language keyword used to define a floating point number variable. The float data type, which is thirty-two bits long, can hold values in the range of ¬-3.402823 x 10 38 to around 3.402823 x 10 38. This type of value is called a single-precision floating-point number.