From the major issues of machine learning and data mining is Feature Selection, picking only the pertinent features to further emphasize the learning search. One relaxed category of feature selection is feature ranking which orders the features according to their relevance. However, numerous feature selection techniques are still rarely used in unsupervised tasks as well as for incrementally clustering continuously emerging mixed data streams with added features. This paper provides a better insight into one particular machine learning algorithm: the Unsupervised Incremental Attribute Learning based on k-prototypes with further analysis on the impact of feature ranking on Incremental K-prototypes. Our approach firstly selects the most relevant features from the new emerging ones with ranking them. Then, a proposed Incremental K-prototypes algorithm is applied on this subset of important features. To do so, two techniques have been proposed. One is the mRMR Feature Selection technique, designed to determine the smallest relevant subset of features with standing for minimum Redundancy and Maximum Relevance. Two is sorting the added features with respect to their calculated variances. The comparative study between the different orders, batch k-prototypes and well-known method reveals that our method is skilled and efficient according to the run time, the Sum of Squared Error and the Davies-Bouldin index evaluation measures.