Herbaceous plants are gaining attention in the pharmaceutical industry due to having less harmful effects reactions and cheaper than modern medicine. However, the diversity of the Ethiopian plant’s species difficulties the identification of herbaceous plants from non-herbaceous ones. And this problem increases the rate of extinction of herbal plants every year. To overcome this problem, in this paper, an efficient approach has been proposed to localize every clearly visible object or region of interest from an image, using less memory and computing power. For object detection we have processed every input image to overcome several complexities, which are the main limitations to achieve better result, such as overlap between multiple objects, noise in the image background, poor resolution etc. We have also implemented an improved Convolutional Neural Network based classification or recognition algorithm which has proved to provide better performance than baseline works. Combining these two detection and recognition approaches, we have developed an Ethiopian Herbaceous Leaf Detection and Recognition (Ethio-Herb) model that is very capable regardless of different limitations such as high and poor image quality, complex background or lightening condition, different leaves of same shape and color, multiple overlapped leaves, existence of non-leaf object in the image and the variety in size, shape, angel and feature of leaf. The experimental results on a custom dataset with 4500 real-world sample leaves and 90 kinds of leaves showed that our proposed model is capable of detecting and recognizing Ethiopian herbaceous leaves from an image with a better accuracy and average precision rate of about 0.997.