Color and texture are the important features used in Content-based image retrieval (CBIR) systems. CBIR is a process that searches and retrieves images from large image databases. To perform this operation, CBIR requires color, texture and shape features of images. In this paper, color and texture features of images are considered. First order statistics and run-length characteristics of images corresponding to color and texture features are extracted. Then, different weights are assigned to each feature which is represented as weighted features. The primary goal is to determine which feature dominates in image retrieval. Similarity measurement between query image and database images are done using sum of absolute difference (SAD), Sum of squared absolute difference (SSAD) and Euclidean distance(ED). It is found experimentally that the weighted features using Euclidean distance gives better performance.