Particle　swarm　optimization　(PSO)　is　widely　used　to　solve　various　optimization　problems,　such　as　robotics　visual　perception　and　intelligent　control　under　uncertainties,　due　to　its　simple　rules　and　easy　implementation.　However,　the　PSO　has　premature　convergence　in　the　optimization　process,　which　will　lead　to　inaccurate　problems　such　as　uncertainties　of　the　control　system.　To　improve　PSOs　performance,　a　self-regulating　particle　swarm　optimization　with　mutation　mechanism　(SRM-PSO)　is　proposed　in　this　paper.　SRM-PSO　combines　the　mutation　mechanism,　self-regulation　and　self-perception　strategy.　The　mutation　mechanism　is　introduced　to　generate　trial　particle　moving　in　different　directions　to　maintain　population　diversity.　Self-regulation　and　self-perception　enable　particles　to　be　updated　in　different　ways　for　fast　exploration　and　intelligent　exploitation.　To　validate　the　effectiveness　of　the　SRM-PSO,　experiments　are　conducted　in　the　CEC2017　test　suite.　The　test　results　indicate　that　SRM-PSO　outperforms　two　related　variants,　and　five　representative　PSO　variants.　Further,　SRM-PSO　is　applied　to　several　real-world　optimization　problems,　which　demonstrates　its　potential　and　competitiveness.