DNA splicing in genomes is a critical and important task to pay attention. Several approaches have been proposed for genic region prediction accuracy in Eukaryotes that mostly rely on digital signal processing techniques with incorporation of indicator sequence methods but an optimal solution is required that could provide the best minimization of 1/f noise and maximization of discrimination between exons and introns. In this paper, we have proposed an enhanced and robust technique for exonic prediction by introducing a novel UTP (University Technology PETRONAS) indicator sequence with denoising target DNA signal using discrete wavelet of third order. We have calculated the percentage improvement in accuracy as compared to existing approaches over traditional (commonly used in literature) datasets PHIX174 (accession number V01128 from location 101 containing 1284 base pairs), gene TUP1 of DNA sequence S.cerevisiae chromosome III (Accession number NC001135) from location 309 and gene F56F11.5 of C elegans (Accession number AF099922) from location 7021. The results obtained from proposed criterion reduced the computational complexity to 75 than Binary indicator sequence method and showed an improvement of 166 as compared to IIR antinoch filter, 133 to Binary indicator sequence method, 130 to filter 2 (multistage filter), 65 to EIIP indicator sequence method and 36 to Complex indicator sequence method. We tested our approach over four other datasets and found results very close to NCBI range.