With the advent of technology, digital signal processing applications are flourishing prominently in space, medical and many commercial related areas. RISC processor is the heart of many high speed applications of embedded and digital signal processing. Floating point representation has prevalent ascendancy over fixed point numbers as it endeavours dynamic range of values. Hence in this paper a high speed MIPS based 32 bit RISC processor with single precision floating point unit for DSP applications is proposed. The inclination of the entire design is towards improving the performance of floating point arithmetic unit so as the performance of the entire RISC processor is ameliorated. The proposed processor is proficient of executing arithmetic, logical, floating point, data transfer, memory, shifting and rotating instructions. The complex multiplication are frequently used in the DSP applications and thus a special instruction for complex multiplication is incorporated. The multiplication engross most of the time, power and area of any operation, on that account the multiplier are reduced in number from four to two as compared to conventional complex multiplication method.