With the rapid growth of various mobile applications and services, the contradiction between the battery capacity and the energy consumption at mobile terminals is becoming increasingly prominent. In addition, the ultra-dense deployment of small base stations (SBSs) in ultra-dense network (UDN) makes network interference more complicated, and servers deployed at the edge of the network are also vulnerable to malicious attacks. In view of this, by jointly optimizing the user equipment (UE) association, cryptographic service assignment, UE power control, and computational resource allocation of UEs and SBSs, the sum of weighted standardized total energy consumption and standardized total security cost is minimized for the multi-task UDN. Specifically, mobile edge computing (MEC) and local computing models are first built for multi-task UDN. Then, a mixed-integer and non-linear optimization problem with minimizing the sum of weighted normalized total energy consumption and normalized total safety cost is formulated, and a joint optimization strategy is designed for it. In such a strategy, an adaptive diversity-guided genetic algorithm (ADGGA) is first executed for a coarse-grained search, and then adaptive particle swarm optimization (APSO) algorithm is run for a fine-grained search. The two-step search procedure is regarded as a further improved hierarchical adaptive search (FIHAS) algorithm. The simulation results show that the proposed algorithm may achieve better system performance than other existing algorithms.