To enhance the convergence speed and solution accuracy of the particle swarm optimization algorithm, a novel improved algorithm, namely the Multi-Strategy Particle Swarm Optimization via Bee Foraging Learning (MSPSO-BFL), is proposed based on the fusion of the particle swarm optimization and artificial bee colony algorithms. Four improvement strategies are separately introduced at different stages of the Bee Foraging Learning-based Particle Swarm Optimization (BFL-PSO). In the initialization stage, a chaotic initialization strategy is adopted to enhance population diversity. In the early iteration phase, an adaptive parameter adjustment mechanism and a global optimum guidance strategy are combined to strengthen the global exploration ability of particles at the beginning of the iteration and accelerate their convergence to the global optimum region in the middle and later stages. In the late iteration phase, Lévy flight with dynamically adjusted step sizes is employed for fine-grained search. In the scout learning phase, an Elite Archive is constructed so that new particles can draw information from it to accelerate convergence. Tested on the CEC2014 benchmark functions, when compared with the original BFL-PSO algorithm, other advanced swarm intelligence algorithms, the proposed MSPSO-BFL algorithm shows significant advantages in solution accuracy and overcomes the poor performance issue of BFL-PSO in dealing with unimodal functions.