Climate variability increasingly threatens agriculture in monsoon-dominated regions, where crop performance depends on both rainfall amount and temporal distribution. This study assessed 26 years (2000–2025) rainfall variability, trends, change points and rainfall state persistence across six agro-climatic zones of West Bengal, India, considering three agricultural seasons: pre-kharif (March–June), kharif (July–October), and rabi (November–February). Seasonal and annual rainfall were analysed using Mann–Kendall and Hamed Rao Kendall tests, Sen’s slope, Pettitt change-point analysis, Innovative Trend Analysis (ITA), rainfall anomalies and LOESS, while rainfall-state dynamics were evaluated using a four-state Markov-chain framework. The Old Alluvial Zone showed the strongest drying signal, with significant declines in pre-kharif, kharif and annual rainfall. ITA confirmed significant declines in kharif and annual rainfall. The Hill Zone exhibited contrasting seasonal trajectories, with increasing kharif rainfall (τ = 0.300; 17.72 mm yr⁻¹; p = 0.038) but declining rabi rainfall (τ = −0.327; −2.17 mm yr⁻¹; p = 0.024), highlighting pronounced seasonal divergence. Markov analysis showed dominant no-rain conditions during rabi, with steady-state probabilities of 0.84–0.92, while kharif heavy-rain probabilities reached 0.33 and 0.30 in the Hill and Terai (Terai-Teesta Alluvial Zone) zones, respectively. The integrated assessment helps identify zones where declining seasonal rainfall, persistent dry conditions or greater monsoon rainfall may have increasing relevance for crop and water management. These rainfall-based insights can support location-specific consideration of supplemental irrigation, moisture conservation, rainwater harvesting, drainage and contingency crop planning.