Picture fuzzy set is an extension of fuzzy sets and intuitionistic sets. It is demonstrated have a wide application in the fact and theoretical. In this paper, we propose some novel similarity measures between picture fuzzy sets. The novel similarity measure is constructed by combining negative functions of each degree membership of picture fuzzy set. This similarity is shown that is better other similarity measures of picture fuzzy sets in some cases. Next, we apply them in several pattern recognition problems. Finally, we apply them to find the fault diagnosis of the steam turbine.