The recognition of handwritten documents, which aims at transforming written text into machine-encoded text, is considered as one of the most challenging problems in the area of character recognition and an open research area. Especially ancient manuscripts, like Ethiopic Geez scripts, chant scripts are different from the modern documents in various ways such as writing style, morphological structure, and writing materials and so on. Geez is one of the ancient languages, which has been used as a liturgical language in Ethiopia. The chant document contains many manuscripts and written using this language contains many unexplored content, which is the base of the current Ethiopic scripts and representation of hymn melodies; however, only few researches have been done on these valuable documents. In this research, the application of ancient handwritten Saint Yared hymn character recognition with Convolutional neural network implementation for the 211 main character set of Ge’ez language is conducted. The training and testing data sets are collected from ancient Ge’ez chant vellum books. In this research, the researchers used various techniques at each phase from digitization to recognition levels. MATLAB image processing is used for experimentations. The iterative Thresholding for binarizing the digitized image, bi-level filtering for noise removal, location detection and horizontal profile for feature extraction methods are found to work very well for the problem of interest. Convolutional Neural Network (CNN) is used for classification. This classifier is trained with 300 pages chant documents. The Classifier is also tested using training datasets. Accordingly, an average recognition rate of 84.6%, 86.8% and 87.2% are registered for oldest chant document and middle age document images, and below 20 years age document image respectively. In recognition rate of 85.4% is registered for average recognition of the system. The performance of the system is greatly affected by the similarity of the shape of chant scripts with other scripts. Invariant to advanced noise detection and removal algorithms should be investigated in the future.