Image retrieval systems now a day's have more focus on searching desired image as a query. Search results are based on different contents of query image. Content based approach is more desirable as it neglects manual annotations of images. Systems aim at search and compare images based on similarity in color contents of query image. Comparison is performed by defining distance similarity measures between color histogram properties. Use of color features as important feature makes results more complete as search results and color contents are not affected by image resize and rotation operations. To extract color features of image we consider on color space, color reduction, color feature extraction process. Our image retrieval application aims at color feature comparison efficiency and accuracy. We aim at reduction in dimensions of local and global color feature which is possible through selection of proper quantization level. To define similarity in feature vector we applied Euclidean distance. Our experimental work shows that local hybrid color histogram produced more accurate image retrieval results than global color moments, color coherence vector and traditional color histogram.