A novel color feature descriptor, Multichannel Distributed Local Pattern (MDLP) is proposed in this manuscript. The MDLP combines the salient features of both local binary and local mesh patterns in the neighborhood. The multi-distance information computed by the MDLP aids in robust extraction of the texture arrangement. Further, MDLP features are extracted for each color channel of an image. The retrieval performance of the MDLP is evaluated on the three benchmark datasets for CBIR, namely Corel-5000, Corel-10000 and MIT-Color Vistex respectively. The proposed technique attains substantial improvement as compared to other state-of- the-art feature descriptors in terms of various evaluation parameters such as ARP and ARR on the respective databases.