The slime mould algorithm (SMA), a revolutionary metaheuristic algorithm with streamlined operations and processes, is frequently utilized to solve optimization issues in various fields. This paper proposed a modified slime mold method (HKTSMA) based on multiple adaptive strategies to ameliorate the convergence speed and capacity to escape local optima. In HKTSMA, the scrambled Halton sequence was utilized to increase population uniformity. By Adjusting the oscillation factor, HKTSMA performs better in controlling the step length and convergence. A novel learning mechanism was proposed based on the k-nearest neighbor clustering method that significantly improved the convergence speed, accuracy, and stability. Then, to increase the probability of escaping the local optima, an enhanced adaptive t-distribution mutation strategy was applied. Simulation experiments were conducted with 32 test functions chosen from 23 commonly used benchmark functions, CEC2019 and CEC2021 test suite and 3 real-world optimization problems. The results demonstrated the effectiveness of each strategy, the superior optimization performance among different optimization algorithms in solving high-dimensional problems and application potential in real-world optimization problems.