Abstract: To address the problem of resource provisioning when the computing node of the cloud fails，a Fault-tolerant Resource Provisioning Method（FRPM）for data-intensive applications was proposed in this paper. Initially，the dataintensive application is modeled as a workflow model，and the faulty task recovery time model and the load balance model were established based on the Virtual Layer 2（VL2）network topology. Subsequently，the Non-dominated Sorting Genetic Algorithm（NSGA-III）was adopted to jointly optimize the makespan of data-intensive applications and the load balance of all computing nodes in the cloud to find effective resource provisioning strategies. Finally，the optimal resource provisioning strategy was obtained based on Simple Additive Weighting（SAW）and the Multiple Criteria Decision Making（MCDM）. The experimental results show that FRPM ensures the shortest makespan of data-intensive applications when the cloud platform fails，and at the same time，it can ensure the load balance of all computing nodes of the cloud.