Data Mining is a technique which is used to discover hidden information from a large database. Frequent item set mining is also an important fundamental problem in data mining. Nowadays, most of the researchers are used association rule mining to find correlation between items and items sets resourcefully. Security is also important problem in data mining. To this end, we propose a transaction splitting based on PFP-growth algorithm and frequent items should keep as secured with the help of cryptography algorithms. PFP-Growth algorithm is advanced to FP-growth algorithm. It consists of both preprocessing phase and mining phase. In the preprocessing phase, we used smart transaction splitting method to improve the utility and tradeoff. In the mining phase, the transformed database and user specified threshold value helps to estimate the number of support computations, so that we can gradually reduce the amount of noise required and the information loss caused by transaction splitting. Using frequent item, we find the global association rules based on association rule mining. In this paper, cryptography technique (AES-Advanced Encryption Standard algorithm) is used to secure the frequent item set. Trusted party should preserve the privacy of individual data while the data is distributed among different sites.