Inter-symbol interference (ISI) limits reliability in diffusion-based molecular communication (MC) channels. We propose RLIM, a family of run-length-limited (RLL) codes that form fixed-size codebooks by minimizing the total number of 1-bits, increasing the per-symbol molecule budget under standard power normalizations and thus improving reliability. We develop a provably optimal linear-time greedy decoder that is equivalent to Viterbi decoding under a deterministic last-wins tie-break and has lower computational complexity; empirically, it outperforms first-wins and random Viterbi variants on RLL baselines. Extensive binomial and particle-tracking simulations show that RLIM achieves lower bit error rate (BER) than classical RLL and other prominent coding schemes across a broad range of scenarios.