This paper , based on normalizing well logging data while drilling and correcting depth into true vertical depth and calculating reservoir parameters and etc , combining the practical ease of mobei oilfield , extracted logging and geological pattern characteristic of target oil - gas formation and geosteering mark formation , and used bp neural network and regressive analysis to create predicting mode of geosteering parameter to build relevant contrast curve ; adopted geometry geosteering method to fix on die drilling direction of bit upper and declination , the position in reservoir , to judge the real drilling case . all finely solved the problem to follow the geological target while drilling for three horizontal well these methods improve the drilling horizontal well ability by using the techniques to follow the geological target while drilling , and then it is convenient and practicable 本文在開展隨鉆測井資料的標(biāo)準(zhǔn)化和斜井校正及儲層參數(shù)解釋與含流體性質(zhì)判釋等工作的基礎(chǔ)上,結(jié)合研究工區(qū)莫北油田的實際情況,提取了目標(biāo)油(氣)層和導(dǎo)向標(biāo)志層的測井地質(zhì)模式特征,并采用bp神經(jīng)網(wǎng)絡(luò)法和回歸分析法建立了地質(zhì)導(dǎo)向參數(shù)的預(yù)測模型、構(gòu)造了相應(yīng)的對比曲線;采用幾何導(dǎo)向法確定鉆頭上下傾鉆進方向及其在目標(biāo)層的位置,以判斷實際鉆進地層情況,很好地解決了研究工區(qū)三口水平井的隨鉆跟蹤地質(zhì)目標(biāo)的問題。