The valuable information is extracted through data mining techniques. Recently, privacy preserving data mining techniques are widely adopted for securing and protecting the information and data. These techniques convert the original dataset into protected dataset through swapping, modification, and deletion functions. This technique works in two steps. In the first step, cloud computing considers a service platform to determine the optimum horizontal partitioning in given data. In this work, K-Means++ algorithm is implemented to determine the horizontal partitioning on the cloud platform without disclosing the cluster centers information. The second steps contain data protection and recover phases. In the second step, noise is incorporated in the database to maintain the privacy and semantic of the data. Moreover, the seed function is used for protecting the original databases. The effectiveness of the proposed technique is evaluated using several benchmark medical datasets. The results are evaluated using encryption time, execution time, accuracy, and f-measure parameters.