Detection studies of fetal abnormalities are essential particularly in the first trimester. These are important for physicious and patients as abnormality can be corrected in the early stage of growth of fetus. Ultrasound techniques give us good opportunity to known about abnormal his such as Anencephaly, Renal anomalies, Cystic Lymphangioma, Cystic Hygroma, Gastroschisis and Fetal megacystis. Firstly the fetal data has to be processed by proper methods such as those used in digital image processing to generate classification of abnormalities.Secondly comparisons with equivalent methods are to be made. After preprocessing using median filter to remove noise, segmentation is done to give qualitative and quantization analysis. This is followed by feature extraction and feature selection using Particle Swarm optimization .Ultimately classification is done by Adaptive Neural Fuzzy Inference System .Comparison of this method with others is carried out to vindicate the efficacy of the proposed technique. The results show that the classification done by the proposed method scores over that of Naive-byes Supports Vector Machine, Linear Discernment Analysis and K-nearest neighbor methods. The proposed method has advantages in detection and classification of the seven abnormalities taken up.